JP2002366728A - Information recommendation system and computer program for it - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information recommendation system by which a user trying to retrieve information on merchandise or services and trying to receive recommendation (recommendation on information on recommended merchandise/ service) can perform retrieval in sufficient consideration to his/her sense of values and can receive recommendation in sufficient consideration to the sense of values, and to provide a computer program for allowing a computer to execute prescribed processing for realizing this. SOLUTION: It is held how users having respective senses of values evaluate the merchandise and service based on the sense of values of the user as a database. Thus, it is possible to recommend merchandise and services which are selected by many persons having the sense of values close to that of the user.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a search for information on goods and services (hereinafter, simply referred to as "goods" in this specification) such as a SW package, a travel package, and restaurant information, and a recommendation of information (recommended). And a computer program used for the system.

[0002]

2. Description of the Related Art Conventionally, in a system for retrieving information on a product and making a recommendation on the information, it has been general to perform a search and a recommendation based on information such as price, quality and quality of the product. .

[0003] Also, clustering (classification) is performed in advance.
And the recommendation of information is made based on the result according to conditions specified by a user using such a system.

[0004]

In the above-mentioned conventional system for searching for information on products and recommending information, search and recommendation are performed based on information such as price, quality and quality of products. , Search with due consideration of the values of the user of the system (for example, the concept that the user depends on when selecting a product, the feeling of expectation of the product selected by the user, etc.) No recommendations were made.

Further, clustering (classification) is performed in advance.
If you want to recommend information based on the results, the classification (clustering)
May not be suitable for individual users. For example, when clustering is performed, when a certain user is located farthest from the center of the cluster, the characteristics of the cluster cannot be said to represent the characteristics of the user. When information recommendation is performed based on the result of such clustering (classification), there is a problem that information on a product desired by a user is not necessarily obtained.

[0006] The present invention fully considers the values of the user (for example, the idea that the user depends on when selecting a product, the feeling of expectation of the product selected by the user, etc.). It is another object of the present invention to provide an information recommendation system capable of searching and recommending information on a product, and a computer program used for the system.

[0007] Further, the present invention provides an information recommendation system in which clustering (classification) is performed in advance, and information recommendation is performed based on a result of the clustering. It is an object of the present invention to provide an information recommendation system capable of recommending information on a suitable product and a computer program used for the system.

[0008]

Means for Solving the Problems In order to solve the above problems, an information recommendation system proposed by the present invention comprises an information obtaining means, an information output means, a storage section, and a processing section. It is characterized in that it is configured as follows.

First, the storage unit stores in advance a product DB in which information on a plurality of products is stored and an information recommendation system provider in advance for each product stored in the product DB. Item DB that stores information about multiple "value items"
And, the information recommendation system provider, in advance for each product stored in the product DB,
For each of the plurality of “value items” provided, the degree of importance of the information recommendation system user is the product-specific value item evaluation result stored for each product stored in the product DB. DB.

[0010] Then, the processing unit is configured to, for a plurality of “value items” stored in the value item DB, included in the information obtained from the information recommendation system user by the information obtaining means, perform the information recommendation. Based on first information indicating how much importance the user of the dating system regards, information on a product corresponding to the first information is stored in the product-specific value item evaluation result DB.
And a product DB.

[0011] Information on the product corresponding to the first information extracted as described above is output via the information output means.

In the above, the information on a plurality of products stored in the product DB includes information on a plurality of products to be searched and recommended. For example, the product DB of the travel package can be configured as shown in FIG. In FIG. 1, symbols such as P1, P2, and P3 in the “product id” column are identifiers for identifying product names.

In the above, the value item DB is, for example, for each product of the travel package in advance,
In the case of a value item DB including a plurality of "value items" provided, the value item DB can be configured as shown in FIG. In FIG. 2, V1 in the “value item id” column,
Symbols such as V2 and V3 are identifiers for identifying value items. By assigning such value items in advance, not only the price, quality, quality, etc. of the product, but also the user's idea of the base upon which the user selects and purchases the product, the expected emotion, etc. Search and recommendation (recommendation of recommended products) that reflect the values of the products.

The product-specific value item evaluation result DB described above is configured, for example, as shown in FIGS. 3A and 3B for each product of the travel package.

FIG. 3A shows the information recommendation system users U1, U2, and U3 of the product represented by the identifier P1 in the product DB shown in FIG. This indicates the degree to which the value item is regarded as important. FIG. 3B shows the information recommendation system users U1 and U for the product represented by the identifier P2 in the product DB shown in FIG.
2, U3 indicates the degree to which the value items indicated by identifiers V1 to V7 are regarded as important.

In the product-specific value item evaluation result DB shown in FIGS. 3A and 3B, the value items indicated by identifiers V1 to V7 in FIG. 2 (that is, the information recommendation system provider The degree to which the information recommendation system users U1, U2, and U3 attach importance to each of a plurality of “value items” previously assigned to each product stored in the product DB illustrated in FIG. (E.g. "I value it" or "I don't value it"
(The degree of importance of the value item) (The degree of importance is determined by the information recommendation system user U1 or the like from the information acquisition unit of the information recommendation system of the present invention, for example, an input unit such as a computer keyboard,
The information is included in information provided to an information obtaining unit of a computer that obtains information provided through a wired or wireless network such as the Internet, etc.). It is displayed after being converted to the numerical value.

Further, the information recommendation system users U1, U2, U3 use the products P1, P2 and provide information such as "satisfied" and "not satisfied" for each of the value items V1 to V7. Figure 3
(A), (b) Like the product-specific value item evaluation result DB shown in the figure, this information is converted into a numerical value of “satisfaction” and displayed based on a predetermined calculation formula.

In the information recommendation system of the present invention, for each of the plurality of products held in the product DB, the value recommendation system of the present invention is used for each of the value items held in the value item DB. The "degree of importance" indicated by the person, that is, the evaluation is
It can be expressed as "important" or "not important", and the degree of importance is expressed by ranking and leveling, such as 1 to 4 levels, 1 to 5 levels, etc. You can also. Further, the expressions “important” and “not important” and the expressions given by ranking and leveling are expressed by using a predetermined calculation formula shown in FIG. ), (B) As shown in the product-specific value item evaluation result DB shown in FIG. It is advantageous to perform a more specific evaluation by displaying a numerical value such as “expectation”.

In the information recommendation system of the present invention, the processing unit is stored in the value item DB (FIG. 2) included in the information obtained from the information recommendation system user by the information obtaining means. For a plurality of "value items" V1 to V7, based on the first information indicating the degree of importance of the information recommendation system user, information on the product corresponding to the first information is set as the product Different value item evaluation result D
B (FIGS. 3A and 3B) and a product DB (FIG. 1) are extracted.

Here, the first information indicating how much importance the information recommendation system user places on the plurality of “value items” V1 to V7 stored in the value item DB (FIG. 2) is For example, in FIG.
The information recommendation system user “emphasizes” the value items represented by 1, V2, V3, etc.
This is an evaluation that expresses the recognition that "it is not important."

For example, a user who has accessed the information recommendation system of the present invention has a “value item” V
The evaluation tendency of “emphasizing” and “not emphasizing” expressed for the items 1 to V7 is indicated by the information recommendation system user U1 in the product-specific value item evaluation result DB illustrated in FIGS. , U2, and U3 are determined to be closer to each other, and information on the product determined to have a similar tendency is stored in a product DB (FIG. 1).
And extract it.

As a result, a user who has accessed the information recommendation system of the present invention can search for a product that reflects his or her own values (the way of thinking, feelings, etc., upon which to select a product) more strongly. And receive recommendations.

The information recommendation system according to the present invention can be realized by causing a CPU of a computer to execute a program stored in a storage medium or the like in advance or a program obtained by the computer via a network such as the Internet. .

In this case, for example, a computer program that causes a computer including an information acquisition unit, an information output unit, a storage unit, and a processing unit to perform the following processing can be adopted as the computer program. .

That is, provided in the storage unit,
A plurality of items stored in a value item DB in which information on a plurality of “value items” assigned in advance to each product is stored in a product DB in which information on a plurality of products is stored. The first information included in the information obtained by the information obtaining means, which represents to what extent the person using the information recommendation system of the present invention attaches importance (evaluation) to the “value item” of the present invention. The processing unit is provided with information on a product corresponding to the first information in the storage unit, and is provided in advance to each product stored in the product DB. For each of the plurality of “value items”, the degree to which the person who uses the information recommendation system of the present invention regards the importance as the value stored in the product DB A process of extracting using the product-specific value item evaluation result DB accumulated for each product and the product DB, and information on a product corresponding to the extracted first information, via the information output means. Is a computer program that causes a computer to perform output processing.

In the information recommendation system of the present invention, the processing unit further provides the information recommendation system with the first information included in the information obtained from the information recommendation system user by the information obtaining means. The user collects and analyzes in advance each of the plurality of “value items” given to each product stored in the product DB, and compiles the information generated by the tallying and analysis. It can be configured to perform processing to accumulate in the product-specific value item evaluation result DB.

By adopting such a configuration, users of the information recommendation system can access each of the products P1P to P3P stored in the product DB (FIG. 1).
For P10, a plurality of “value items” V in the value item DB
Information (evaluation) indicating that each of the items 1 to V7 is "important" or "not important" is totaled and analyzed, and the product-specific value item evaluation result DB (FIG. 3A,
(B)). Therefore, the contents of the product-specific value item evaluation result DB are enriched each time the information recommendation system user accesses and provides the first information. The user who has accessed the information recommendation system of the present invention selects his or her own values (products) by performing search and recommendations using the product-specific value item evaluation result DB as described above. Products and products that more strongly reflect the way of thinking and emotions on which they are based, and receive recommendations.

The processing in such a processing unit is also performed in the computer program described above, and further in the processing unit.
The first information included in the information acquired by the information acquiring means is tabulated and analyzed for each of the plurality of “value items” previously assigned to each product stored in the product DB. Then, it can be realized by causing a computer of a computer to execute a computer program for performing processing for accumulating the information generated by the tallying and analysis in the product-specific value item evaluation result DB.

In the information recommendation system according to the present invention, the storage unit further stores the degree of importance of each information recommendation system user for each of the plurality of value items in the value item DB. It can be configured to have an area for the user-specific value item evaluation result DB. Further, the processing unit may further include the first information included in the information obtained from the information recommendation system user by the information obtaining means, for each of the plurality of value items in the value item DB, for each information recommendation. Tallying and analyzing the degree of emphasis for each user of the dating system,
The information generated by the analysis is stored in the user value item evaluation result DB area for the user value item evaluation result DB.
It can be configured to perform the process of holding as.

FIGS. 4A, 4B, and 4C show an example of a user-specific value item evaluation result DB.

FIG. 4A shows that each of the products P1P to P10 stored in the product DB (FIG. 1) which the information recommendation system user U1 has accessed.
From the information indicating that each of the plurality of “value items” V1 to V7 is “emphasized” or “not important”, importance is placed on each of the plurality of value items V1 to V7 in the value item DB. Tally, analyze,
The results of the aggregation and analysis are displayed. FIG.
(B) is the information recommendation system user U2
FIG. 4C shows the results of the tallying and analysis in the case of the information recommendation system user U3.

The processing in such a processing unit is also performed in the computer program described above, and further in the processing unit.
Tallying the first information included in the information obtained by the information obtaining means, for each of the plurality of value items in the value item DB, for the degree of importance of each person who provided the first information, Analyzing, totaling, and information generated by the analysis, for each of a plurality of value items in the value item DB, information on the degree of importance of each person who provided the first information is stored for each user This can be realized by causing a CPU of a computer to execute a computer program for performing processing to be held in the value item evaluation result DB area.

[0033] In the information recommendation system of the present invention, the storage unit further includes a plurality of groups for dividing the information recommendation system users into a plurality of groups according to the first information. A group DB storing information on the degree of importance of each of the plurality of “value items” of the information recommendation system user divided into each group of the plurality of groups; It can be configured to include a recommendation DB area including information, second group information, third group information, fourth group information, and fifth group information.

The processing unit further uses the group DB according to the first information included in the information obtained from the information recommendation system user by the information obtaining means, and The information recommendation system user who has provided the information is divided into any one of a plurality of groups in the group DB, and according to the first information, the product-specific value item evaluation result DB And the first group information, the second group information, the third group information, the fourth group information, and the fifth group information using the user-specific value item evaluation result DB. Generate information consisting of
It can be configured to include a value item segmentation processing unit that performs a process of storing these as a recommendation DB in the recommendation DB area.

Here, the first group information is, when the information recommendation system user is divided into a plurality of groups according to the first information, the divided group and the first information. Associates each information recommendation system user that has provided, and indicates which group of the plurality of groups each information recommendation system user that has provided the first information respectively. Information, for example,
This is displayed as shown in FIG. FIG.
(A) In the information recommendation system user U
1 is grouped into group 3, information recommendation system user U2 is grouped into group 1, and information recommendation system user U3 is grouped into group 5.

The first information (ie, value item DB)
Is information indicating how much importance the information recommendation system user places on a plurality of "value items" stored in the "Valid Item". For example, in FIG.
The user of the information recommendation system “values” the value items represented by 1, V2, V3, etc.,
A statement expressing recognition (evaluation) of "not emphasizing". ), The information recommendation system user is divided into a plurality of groups. For example, in the group DB, the information recommendation system user who attaches importance to the value items V1 and V3 for the product P1 is group 1 And the information recommendation system users who value the value item V2 with respect to the product P2 are grouped into a plurality of groups such as group 2 and the like. The information relating to the degree of importance of each of the plurality of “value items” is stored, and can be performed using this group DB.

Further, the second group information includes information of the information recommendation system users belonging to each group when the information recommendation system users are divided into a plurality of groups according to the first information. , And information on the average value of the degree of importance for each of the plurality of “value items”, for example, as shown in FIG.
It is displayed as shown.

FIG. 5B shows a group 1 (identifier G1).
In each of the groups divided into group 5 (identifier G5), the average value of the degree of importance of each of the value items represented by the identifiers V1 to V7 is converted into a numerical value by a predetermined calculation formula and displayed. Is what is being done.

Further, the third group information is stored in the product DB in each group when the information recommendation system user is divided into a plurality of groups according to the first information. This is information on the average value of the degree of importance for each of the plurality of “value items” for each product, and is displayed, for example, as shown in FIG. 5C.

FIG. 5C shows a group 1 (identifier G1).
Each product (P1, P2,...) Stored in the product DB
.., P10), the average value of the degree of importance for each of the plurality of “value items” (identifiers V1 to V7) is converted into a numerical value of the degree of expectation by a predetermined formula and displayed. Is what it is.

The users of the information recommendation system grouped into group 1 (identifier G1) use the products P1, P2,..., P10, and state “satisfied” for each of the value items V1 to V7. If information such as "not satisfied" is provided, this information is converted into a numerical value of "satisfaction degree" based on a predetermined calculation formula, and the information shown in FIG. As described above, for each product stored in the product DB in each group, it can be displayed as the degree of satisfaction for each of the plurality of “value items”.

Next, the fourth group information includes, for each of the plurality of “value items” when the information recommendation system users are divided into a plurality of groups according to the first information, Each information recommendation system user belonging to each group is information on the degree of importance given to each user. For example, FIG.
It is configured like the user-specific value item evaluation result DB shown in (b) and (c).

Further, the fifth group information includes, for each of the plurality of “value items”, when the information recommendation system users are divided into a plurality of groups according to the first information. The degree of importance of each information recommendation system user belonging to the group is the information stored for each product stored in the product DB, for example, as shown in FIGS. 3 (a) and 3 (b).
Is constructed like a product-specific value item evaluation result DB shown in FIG.

Such value item segmentation processing in the processing unit can be realized by causing the CPU of the computer to execute a computer program that causes the processing unit to perform the following processing in the computer program described above.

That is, using the group DB provided in the storage unit, in accordance with the first information included in the information obtained by the information obtaining means, the product that has provided the first information. The buyer is divided into any one of a plurality of groups, the first group information is generated, and the product-specific value item evaluation result DB and the user are generated in accordance with the first information. Using the different value item evaluation result DB, the second group information, the third group information, the fourth group information, and the fifth group information are respectively generated, and information including these is stored in the storage unit. The computer program for performing the process of storing the information in the recommendation DB area is stored in the C of the computer.
It can be realized by being executed by a PU.

Further, in the information recommendation system according to the present invention, the processing unit may be configured such that the information recommendation system provider provides the plurality of products previously assigned to each product stored in the product DB. For each of the "value items", according to the second information, which is included in the information acquired from the information recommendation system user by the information acquisition means and indicates the degree of importance of the information recommendation system user Then, a process of extracting information on a product corresponding to the second information can be performed using the recommendation DB and the product DB.

Here, the second information is stored in a value item DB (FIG. 2) which is previously assigned to each of the products P1 to P10 stored in the product DB (FIG. 1). Information indicating how much importance the user of the information recommendation system regards a plurality of "value items" V1 to V7, and is represented by, for example, identifiers V1, V2, V3, etc. in FIG. With respect to the value items, the information recommendation system user has decided that “the degree of importance: 1 (not at all important)”, “the degree of importance: 2 (not very important)”, “ Degree: 3 (Slightly emphasized) "," Degree of importance: 4
(Very important) ", and the degree of importance is evaluated by ranking or leveling.

The above-described processing performed based on the second information is performed when the information recommendation system users belonging to each group are divided into a plurality of “values” when the information recommendation system users are divided into a plurality of groups. The second group information, which is information on the average value of the degree of importance (evaluation) for each of the “items”,
In comparison with the second information included in the information obtained from the information recommendation system user by the information acquisition means, those who have accessed the information recommendation system, which group of a plurality of groups It is determined whether the product has a similar tendency, and then the product D in the group determined to have a similar tendency to those who have accessed the information recommendation system.
Third group information that is information on the average value of the degree of importance (evaluation) for each of the plurality of “value items” V1 to V7 for each of the products P1 to P10 stored in B, and the group concerned The fourth group information, which is information on the degree (evaluation) of each information recommendation system user belonging to each of the plurality of “value items” to be regarded as important, and each information recommendation system user belonging to the group The degree of importance (evaluation) of each of the plurality of “value items” is determined by referring to the fifth group information, which is information stored for each product stored in the product DB. Information about the product corresponding to the second information is extracted.

The information about the product corresponding to the second information extracted in this way is output through the information output means. The information is based on the values (multiple values) of the information recommendation system user. “Value Items”
The degree of importance (evaluation) of each of V1 to V7 and a plurality of “value items” for each of products P1 to P10
V1 to V7 are extracted with reference to the evaluation results of people who have close importance. As a result, a user who has accessed the information recommendation system of the present invention is more likely to reflect his or her own values (thoughts, feelings, etc., upon which to select a product), You can search and get recommendations.

The processing in such a processing unit is also performed in the computer program described above, and further in the processing unit.
The present invention can be realized by causing a CPU of a computer to execute a computer program for performing the following processing.

That is, for each of the plurality of “value items”, the importance of the information recommendation system user included in the information acquired from the information recommendation system user by the information acquisition means. In accordance with the second information indicating, using the recommendation DB and the product DB, performs a process of extracting information about a product corresponding to the second information, and performs the process of extracting the extracted second The present invention can be implemented by causing a CPU of a computer to execute a computer program for performing a process of outputting information on a product corresponding to the information via the information output unit.

Further, in the information recommendation system of the present invention, the information recommendation system provider may provide the plurality of “value items” previously assigned to each product stored in the product DB. For each of the above, included in the information obtained from the information recommendation system user by the information acquisition means, according to the second information indicating how much importance the information recommendation system user regards The unit uses the recommendation DB and the product DB to perform a process of extracting information on a product corresponding to the second information, instead of or in addition to the above-described configuration. , Can also be configured as follows.

That is, the storage unit further stores, for each of the plurality of “value items” previously given by the information recommendation system provider to each product stored in the product DB, According to the second information, which is included in the information obtained from the information recommendation system user by the information obtaining means and indicates the degree of importance of the information recommendation system user, the second information The information recommendation system has provided a grouping DB in which information on a grouping method for grouping users is stored, and an area for a user unit value model DB. The processing unit further includes: The processing is performed.

In this process, first, the information recommendation system provider provides a plurality of “value items” previously assigned to the products P1 to P10 stored in the product DB (FIG. 1). For each of V1 to V7, the second information, which is included in the information obtained from the information recommendation system user by the information obtaining means and indicates the degree of importance of the information recommendation system user, Accordingly, a process of forming a new group using the grouping DB so that the information recommendation system user who has provided the second information is located at the center of the group.

Second, by using the product-specific value item evaluation result DB and the user-specific value item evaluation result DB, the information of the information recommendation system users belonging to the new group formed as described above. And, for example, FIG.
(A) As shown in the figure, the sixth group information to be displayed is generated, and each of the plurality of “value items” of the information recommendation system user belonging to the new group formed as described above. Generates seventh group information related to the average value of the degree of importance of the plurality of "value items" of each product stored in the product DB in the new group formed as described above. Generate eighth group information about the average value of the degree of emphasis on each, and each information recommendation system user belonging to the new group formed as described above, for each of the plurality of "value items" , Generating ninth group information relating to the degree of importance, and each information recommendation belonging to the new group formed as described above. The tenth group information in which the degree of importance of each of the plurality of “value items” by the application system user is summarized for each product stored in the product DB is generated, and The information composed of the sixth group information, the seventh group information, the eighth group information, the ninth group information, and the tenth group information generated as described above is stored in the user unit value model DB. As a user unit value model DB in the storage unit
This is a process of storing the information in the area for use.

Then, in accordance with the second information, the user unit value model DB and the product DB are used to extract information on a product corresponding to the second information.

As a result, information on the product corresponding to the extracted second information is output via the information output means.

According to such a configuration, the information recommendation system provider provides a plurality of "value items" V1 previously assigned to the products P1 to P10 stored in the product DB (FIG. 1). To V7 in accordance with the second information, which is included in the information obtained from the information recommendation system user by the information obtaining means and indicates the degree of importance of the information recommendation system user. Using the grouping DB, a new group is formed such that the information recommendation system user who has provided the second information becomes the center of the group. Therefore, when the classification (clustering) is performed, a user who accesses the system of the present invention for information recommendation is placed at the farthest position of the cluster, and the characteristic of the cluster is that user. Therefore, even if the recommendation of the service / product based on the cluster is performed, the situation that the service / product desired by the user is not necessarily recommended can be prevented beforehand.

In addition, according to the above-mentioned second information, the information recommendation system user who has provided the second information by using the grouping DB is newly set to be located at the center of the group. The processing for forming a group is, for example, "a grouping method for performing a factor analysis so that a factor contribution rate to a value item evaluation result of an information recommendation system user who has provided the second information is maximized", " Information recommendation system that has provided the second information Grouping method that defines the discriminant function so that the distance to the user's value item evaluation result is minimized, "The information recommendation system that has provided the second information A grouping method that performs cluster analysis so that the distance to the user's value item evaluation result is minimized An execution program, such as allowed to held in the grouping DB, can be performed by selecting one in this.

The processing in such a processing unit is also performed in the computer program described above, and further in the processing unit.
The present invention can be realized by causing a CPU of a computer to execute a computer program for performing the following processing.

First, the information / recommendation system provider provides, for each product stored in the product DB,
Each of a plurality of "value items" provided in advance is included in the information obtained from the information recommendation system user by the information obtaining means,
According to the second information indicating how much importance the information recommendation system user attaches to the information recommendation system, the information recommendation system user relates to a grouping method for grouping the information recommendation system users who have provided the second information. This is a process of forming a new group using the grouping DB in which information is stored, so that the information recommendation system user who has provided the second information is located at the center of the group.

Next, using the product-specific value item evaluation result DB and the user-specific value item evaluation result DB, the information of the information recommendation system users belonging to the new group formed as described above is obtained. A certain sixth group information is generated, and the information recommendation system user belonging to the new group formed as described above, regarding the average value of the degree of importance for each of the plurality of “value items”, Seventh group information is generated, and the product DB in the new group formed as described above.
Generates eighth group information on the average value of the degree of emphasis on each of the plurality of "value items" for each of the products stored in the group, and belongs to the new group formed as described above. For each of the plurality of "value items", the information recommendation system user generates ninth group information relating to the degree of importance, and each information recommendation system belonging to the new group formed as described above. The degree to which the user regards each of the plurality of “value items” as important is generated by generating tenth group information in which each product stored in the product DB is summarized and generated as described above. The sixth group information, the seventh group information, the eighth group information, the ninth group information, and the tenth group information Information consisting of the storage unit as the user unit value model DB user unit value model DB
This is a process of storing the information in the area for use.

Then, the information recommendation system provider provides the plurality of “value items” previously assigned to each product stored in the product DB.
For each of the, included in the information obtained from the information recommendation system user by the information acquisition means, according to the second information indicating how much importance the information recommendation system user regards,
This is a process of extracting information on a product corresponding to the second information using the user unit value model DB and the product DB.

Then, a computer program for performing a process of outputting information on a product corresponding to the second information extracted as a result through the information output means can be realized by a CPU of the computer.

Any of the information recommendation systems of the present invention described above may further have the following configuration.

That is, the storage unit further stores the plurality of “value items” previously assigned by the information recommendation system provider to each product stored in the product DB, and The information recommendation system provider has a concept dictionary DB in which a plurality of predetermined “concepts” are stored in advance in correspondence with the “value items” of A word included in a sentence included in the information obtained from the user of the dating system and information relating to the correspondence between the word and the plurality of predetermined “concepts” are accumulated. And a word-concept analysis DB, and an area for a sentence-concept DB described below.

Then, the processing unit further divides the sentence included in the information obtained from the information recommendation system user by the information obtaining means into words and uses the word concept analysis DB to perform the division. Analyzing the relationship between the selected word and the plurality of “concepts”, and associating a sentence containing a word having a relationship corresponding to the plurality of “concepts” with the corresponding “concept” In addition, a text-concept DB is stored in the text-concept DB area of the storage unit as a text-concept DB, and a “concept” corresponding to the plurality of “value items” and / or a word having a relation corresponding to the “concept” is stored. Is included in the concept dictionary DB,
The extraction process is performed using the sentence-concept DB.

The information extracted in this way is output from the information output means.

The above-described system is effective in an environment where a so-called chat or IRC (Internet Relay Chat) is performed. That is, when the information recommendation system provider of the present invention implements the information recommendation system of the present invention on a server computer that can be connected to a network such as the Internet, the information recommendation system of the present invention is used. The user (user) does not or does not input, as information, the evaluations (eg, “important” or “not important”) for a plurality of value items held in the value item DB. In addition to the above, when free text is provided by chat or the like, the above-described system is effective.

A sentence input by chat is generally composed without regard to a plurality of value items held in the value item DB. However, according to the system of the present invention, the concept dictionary DB and the Using the sentence-concept DB, it is possible to associate words included in a free sentence input in the chat with a plurality of value items in the value item DB. Then, by using a sentence-concept DB in which this association is held as data,
The user extracts a sentence containing a word of a concept related to a value item that the user considers important, from a free sentence input by chat held in the information recommendation system of the present invention. You can read it.

Therefore, the information recommendation system of the present invention is used to evaluate a plurality of value items held in the value item DB, such as “important” and “not important”. As described above, the user who provides and inputs the information to the information acquisition unit can extract and read a sentence including a word of a concept related to the value item that the user attaches importance to as described above. By this, information retrieval by the information recommendation system of the present invention, whether or not the recommendation is proceeding reflecting its own values, read the free text that others are input by chat While confirming, the information recommendation system of the present invention can be used.

In the above-described system configuration, in the above-described computer program, the processing unit further divides a sentence included in the information obtained from the information recommendation system user by the information obtaining means into words, Information recommendation system provider,
For each product stored in the product DB, in advance, the information recommendation system provider associates the plurality of “value items” with the plurality of “value items” and A word in which information relating to the correspondence between a plurality of “concepts” in the concept dictionary DB in which a plurality of defined “concepts” are stored and words included in the sentence is stored. Concept analysis DB
Using the divided words and the plurality of “concepts”
Is analyzed, and a sentence including a related word corresponding to the plurality of “concepts” is associated with the corresponding related “concept”, and the sentence-concept DB in the storage unit is associated with the sentence.
And a sentence including a “concept” corresponding to the plurality of “value items” and a related word corresponding to the “concept” are stored in the concept dictionary DB and the sentence-concept. This can be realized by causing a CPU of a computer to execute a computer program for performing a process of extracting using a DB.

In addition to the above system configuration, a person (user) who uses the information recommendation system of the present invention evaluates a plurality of value items held in the value item DB (“emphasis is placed on the value items”). ) Or "don't value" as information, so that you can search for products and receive recommendations only by providing and inputting free sentences by chat or the like. You can also.

That is, in the computer program described above, the processing unit further divides the document included in the information obtained from the information recommendation system user by the information obtaining means into words, To analyze the relationship between the divided words and the “concepts” that are predetermined corresponding to a plurality of value items in the value item DB. "Concept" and value item D
Analyze the corresponding relationship with the value items in B, and use the value item evaluation results DB and recommendation DB for the users and products recommended to the group who value the value items derived as a result. The CPU of the computer executes a computer program for extracting and outputting the extracted information through the information output means, and the extracted information has been provided and input by the free text by chat or the like. The system configuration includes a processing unit that executes a processing operation of providing the processing to a user and a storage unit.

With such a configuration, the user of the information recommendation system “imports” or “does not attach importance” to a plurality of value items in the value item DB.
It is possible to receive a recommendation of a product recommended to a person who evaluates a value item close to oneself simply by performing so-called chatting and inputting a free sentence without performing evaluation such as evaluation.

In the above-described system configuration that can use the information recommendation system of the present invention while reading and confirming the free text input by another person through chat, the following system configuration is adopted. Thus, the information recommendation system of the present invention can be realized.

That is, the storage unit further stores the plurality of value items previously assigned by the information recommendation system provider to each product stored in the product DB, and the plurality of value items. According to the item,
The information recommendation system provider provides a concept dictionary DB in which a plurality of predetermined concepts are stored in advance, and a sentence corresponding to “concepts” stored in the concept dictionary DB. Multiple "conceptual patterns" that divide stored sentences into "character strings with parts of speech"
In the "concept pattern", the registered "part-of-speech appearance" which is the "character string with part-of-speech" information in which only the part-of-speech ignoring the words for the part-of-speech other than the keyword predetermined by the information recommendation system provider And a second word concept analysis DB in which "patterns" are stored, and a second sentence-concept DB area.

The processing unit further analyzes the sentence included in the information obtained from the information recommendation system user by the information obtaining means by dividing the sentence into a “character string with a part of speech” which is a character string with a part of speech. In the analysis "character string with part-of-speech" in the analysis "character string with part-of-speech", analysis "character string with part-of-speech" information in which only the part-of-speech is ignored for parts of speech other than the keyword, "part-of-speech appearance pattern"
By comparing the analysis “part-of-speech appearance pattern” with the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB, the registered “part-of-speech appearance pattern” is similar to the analysis “part-of-speech appearance pattern”. A search is performed to determine whether a “part-of-speech appearance pattern” exists, and a registered “part-of-speech appearance pattern” similar to the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB exists. The analysis “character string with part-of-speech” and the sentence related to the analysis “part-of-speech appearance pattern” that can be confirmed to be performed correspond to the “concept” corresponding to the “concept pattern” related to the registered “part-of-speech appearance pattern”. In the second sentence-concept DB area of the storage unit, the second sentence-concept DB is held as the second sentence-concept DB, and the concept corresponding to the plurality of value items and / or the concept is associated with the concept. "concept The text of the turn ", and the concept dictionary DB, the second sentence - performs a process of extracting with a concept DB.

The information extracted in this way is output from the information output means.

In this system configuration, in the above-described computer program, the processing section is a character string with a part of speech attached to the text included in the information obtained from the information recommendation system user by the information obtaining means. The "character string with part of speech" is divided into "character string with part of speech" and analyzed as "character string with part of speech". And the information recommendation system provider provides the plurality of value items that have been previously assigned to each product stored in the product DB, along with the analysis “part of speech appearance pattern” that is the information. The information recommendation system provider accumulates a plurality of predetermined concepts in correspondence with multiple value items. Concept dictionary DB and a plurality of “concepts” obtained by dividing sentences corresponding to “concepts” stored in the concept dictionary DB and stored in the sentence-concept DB into “character strings with parts of speech” In the “concept pattern” and the “concept pattern”, the registered word “part-of-speech-attached character string” information in which only the part of speech ignoring the words is registered for the parts of speech other than the keywords predetermined by the information recommendation system provider Using the second word concept analysis DB in which the "part of speech appearance pattern" is stored, the analysis "part of speech appearance pattern" and the registered "part of speech appearance pattern" stored in the second word concept analysis DB are used. To find out whether there is a registered “part of speech appearance pattern” similar to the analysis “part of speech appearance pattern”, and store it in the second word concept analysis DB. The analysis "character string with part-of-speech" based on the above-mentioned analysis "part-of-speech appearance pattern" in which it was possible to confirm that a similar registered "part-of-speech appearance pattern" exists in the registered "part-of-speech appearance pattern" And the sentence are associated with the “concept” corresponding to the “concept pattern” related to the registered “part-of-speech appearance pattern”, and the second sentence-concept is stored in the second sentence-concept DB area of the storage unit. The concept dictionary DB and the second sentence-concept DB are used to store a sentence relating to a concept corresponding to the plurality of value items and / or a “concept pattern” associated with the concept. It can be realized by causing a CPU of a computer to execute a computer program for performing a process of extracting by using a computer.

With such a system configuration and a computer program, when a user provides and inputs a free description sentence by chat or the like, the concept dictionary DB and the second word concept analysis DB are used. Then, the free text input in the chat is analyzed into the analysis “part-of-speech appearance pattern”, and the “conceptual pattern” related to the “part-of-speech appearance pattern” and the value item DB are analyzed based on the analysis “part-of-speech appearance pattern”. Multiple value items V1 to V7
Are associated with each other, and this association is held in the second sentence-concept DB held as data. And
In the free text input by the chat held in the information recommendation system of the present invention, the user can use the “concept pattern” and the “part-of-speech appearance pattern” related to the value item that the user emphasizes. Can be read from the second sentence-concept DB. Therefore, in the "concept pattern" in which the free description is divided into "character strings with part-of-speech", and in the "character strings with part-of-speech" which is this "concept pattern", words are ignored for parts of speech other than keywords (nouns). Since the free description is selected based on the correspondence between the "part of speech appearance pattern" in which only the part of speech has been changed and the free description concept, the words and concepts that the information recommendation system provider of the present invention did not recognize were selected. Even if there is a corresponding relationship with the free description sentence, the corresponding free description is selected from the “character string with part of speech (that is, concept pattern)” and “part of speech appearance pattern” of the free description sent by the user. , The user can read it. This allows the user to search for and recommend information using the information recommendation system of the present invention.
Use the information recommendation system of the present invention while reading the free texts that other people are typing by chatting to see if their values are being reflected and proceeding more effectively. be able to.

In the system configuration described above, the processing unit further converts the text included in the information obtained from the information recommendation system user by the information obtaining means into a character string with a part of speech, In the analysis "character string with part-of-speech" in the analysis "character string with part-of-speech", the "part-of-speech string" in which only the parts of speech except for the keyword are ignored
The analysis is the analysis “part of speech appearance pattern”, and the analysis “part of speech appearance pattern” is compared with the registered “part of speech appearance pattern” stored in the second word concept analysis DB.
The registered "part-of-speech appearance pattern" that is similar to the analysis "part-of-speech appearance pattern" is extracted, and the extracted keywords in the registered "part-of-speech appearance pattern" are compared with the keywords in the analysis "part-of-speech appearance pattern". When the two are different, the analysis “character string with part-of-speech” according to the analysis “part-of-speech appearance pattern” is performed according to the registered “part-of-speech appearance pattern” extracted in the second word concept analysis DB. It can be configured to perform a processing operation of registering in the second word concept analysis DB as a new “concept pattern” associated with the “concept” with which the “concept pattern” is associated.

Thus, the contents of the second word concept analysis DB described above are automatically created, and the contents are enriched. That is, the “concept pattern” or “keyword” that was not initially recognized by the information recommendation system provider of the present invention is used in the second word concept analysis DB.
It is automatically accumulated inside and enriched.

In this case, the processing unit further compares the keyword in the extracted registered “part-of-speech appearance pattern” with the keyword in the analysis “part-of-speech appearance pattern”. Even if the "concept pattern" stored in the second word concept analysis DB corresponding to the registered "part-of-speech appearance pattern" is used, the frequency of use of the "concept pattern" in the past is analyzed, and the use frequency is determined in advance. The analysis “character string with part-of-speech” pertaining to the analysis “part-of-speech appearance pattern” pertains only to the registered “part-of-speech appearance pattern” extracted in the second word concept analysis DB only when the threshold value exceeds It can be configured to perform processing for registering in the second word concept analysis DB as a new “concept pattern” associated with the “concept” corresponding to the “concept pattern”. That.

By adopting such a system configuration, the process of accumulating and enriching “concept patterns” and “keywords” in the word concept analysis DB, and enriching the second sentence-concept DB In the process, the more frequently used "concept patterns" and "keywords" are accumulated, so that the reliability of the information recommendation system of the present invention is increased for the user. In addition, thereby, the information recommendation system of the present invention is automatically enriched with contents that are easier for the user to use.

In the processing by these system configurations, in the above-described computer program, the processing section further includes a sentence included in the information obtained from the information recommendation system user by the information obtaining means, in a character with a part of speech. In the analysis "character string with part-of-speech", it is divided into "character string with part-of-speech", and in the analysis "character string with part-of-speech", the part-of-speech state in which only parts of speech except for the keyword are ignored The analysis “part-of-speech appearance pattern”, which is the “attached character string” information, is compared with the analysis “part-of-speech appearance pattern” and the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB. A registered “part-of-speech appearance pattern” that is similar to the “part-of-speech appearance pattern” is extracted, and a keyword in the extracted registered “part-of-speech appearance pattern” Analysis comparing the keyword of the "part of speech appearance pattern" in, when they are different, the analysis the analysis according to the "part of speech appearance pattern""String with part of speech"
As a new “concept pattern” associated with the “concept” associated with the “concept pattern” related to the registered “part-of-speech appearance pattern” extracted in the second word concept analysis DB This can be realized by causing a CPU of a computer to execute a computer program for performing a processing operation to be registered in the concept analysis DB.

Further, in the computer program described above, the processing unit compares the extracted keyword in the registered “part of speech appearance pattern” with the keyword in the analyzed “part of speech appearance pattern”. Is different, the frequency of use of the "concept pattern" stored in the second word concept analysis DB in the past corresponding to the registered "part-of-speech appearance pattern" is analyzed. Only when the value exceeds a predetermined threshold value, the analysis “character string with part-of-speech” related to the analysis “part-of-speech appearance pattern” is performed using the registered “part-of-speech” extracted in the second word concept analysis DB. The second word concept analysis DB as a new “concept pattern” associated with the “concept” corresponding to the “concept pattern” related to the “appearance pattern”
Can be realized by causing a CPU of a computer to execute a computer program for performing a process of registering the information.

[0088]

Preferred embodiments of the present invention will be described below with reference to the accompanying drawings.

An information recommendation system according to the present invention can be realized by causing a CPU of a computer to execute a program stored in a storage medium or the like in advance or a program obtained by the computer via a network such as the Internet. The simplest system configuration includes, as shown in FIG. 7, an information acquisition unit 2, an information output unit 3, and a storage unit 4.
And a processing unit 5, and can be realized by a known computer 1 in which these components are communicably connected to each other by a bus 6.

The information acquisition means 2 includes, for example, a keyboard,
The information output unit 3 can be realized by a CRT or a liquid crystal display, and the processing unit 5 is realized by a CPU.

The storage unit 4 is provided with areas for a plurality of databases (referred to as “DB” in the present specification) to be described later. System), GUI
(Graphical user interface), a browser, and the like. Further, an application program (AP) for executing the information recommendation system according to the present invention (search for a service / product and executing a recommendation) and processing, and the computer program of the present invention described above are stored in the storage unit. 4 may be stored from the beginning, or the computer program of the present invention described above may be stored in an external storage medium such as an MD or a flexible disk, and stored in an information storage device such as a disk drive. Acquisition means 2
To execute the information recommendation system according to the present invention (search for a product or execute a recommendation) and process.

Further, the computer 1 shown in FIG. 7 is connected to a wired or wireless network 7 such as the Internet (for example, as shown in FIG. 7, connected to the network 7 by a line means such as an ISDN line 8). Through the network 7, the computer 1 obtains information from the users U1, U2, U3, U4,. Means (not shown) may be configured to make recommendations to the user U1 and the like via the network 7.

In the information recommendation system of the present invention, the storage unit 4 stores a product DB11 in which information on a plurality of products is stored, and a product DB in which the information recommendation system provider stores each product stored in the product DB11. On the other hand, a value item DB 12 in which information on a plurality of “value items” assigned in advance is stored.
And information recommendation system provider, product D
For each product stored in B11,
The degree to which the information recommendation system user U1 or the like attaches importance to each of the plurality of “value items” provided is the product-specific value item evaluation stored for each product stored in the product DB11. Result DB13
And are provided.

Then, the processing unit 5 stores a plurality of values stored in the value item DB 12 included in the information obtained from the information recommendation system user U1 or the like by the information obtaining means according to the above-described computer program of the present invention. Based on the first information indicating the degree of importance of the information recommendation system user U1 and the like regarding the “value item”, information on a product corresponding to the first information is converted into a value item for each product. Evaluation result DB
13 and the product DB 11 are extracted.

The information on the product corresponding to the first information thus extracted is output via the information output means.

In the above description, information on a plurality of products stored in the product DB 11 includes information on a plurality of products and services to be searched and recommended. For example, the product DB 11 of the travel package can be configured as shown in FIG. In FIG. 1, symbols such as P1, P2, and P3 in the “product id” column are:
It is an identifier for identifying the product name (“Pokapoka Hot Spring Tour”, “Ski Zanmai Tour”, “Crab Eating Roll Tour”,...).

[0097] The information recommendation system provider determines that each of the products P1, P2,
Value item DB1 in which information on a plurality of “value items” assigned to P3 and the like is stored in advance.
2 shows, for example, the value of the user of the product or the target information. N. It can be modeled and used based on Sheth's consumption value theory.

According to the consumption value theory of JNSheth, consumer value items are classified into the following five types. 1) Functional value: Value item related to functional, practical, or physical utility 2) Social value: Value item related to utility reminiscent of a specific social group 3) Emotional value: A specific emotional state Value items related to utility to be evoked 4) Cognitive value: Value item related to utility that provides new knowledge and information 5) Conditional value: Value related to conditional utility that is expected to enhance other values under specific circumstances item

In the prior art, when conducting a questionnaire to users, consumers, and service / product purchasers, the focus is on functional values such as price, quality, quality, and the like. I couldn't get the view, and could only extract value from the supply side's point of view. On the other hand, for example, the user's sense of value can be extracted by focusing on the above five items and evaluating the target product or information.

In the case where the information recommendation system of the present invention is realized as a travel package recommendation system, for example, the aforementioned JNSh
Based on eth's consumption value theory, if you set the consumer's value items, V1) you can relax your schedule V2) you can feel rich V3) you can discover a new self V4) you can forget your daily life V5) You can refresh your mind and body V6) You can taste the nostalgia of the old days V7) You can set the value items that make you feel nostalgic.

The value item DB 12 including the plurality of “value items” is as shown in FIG. FIG.
In the middle, symbols such as V1, V2, and V3 in the “value item id” column are identifiers for identifying value items.

Since the value items V1 to V7 are assigned in advance, not only the price, quality, quality, etc. of the product, but also the ideas and expectations based on which the user selects and purchases the product. This makes it possible to search for products and make recommendations (recommendation of recommended products) that reflect the user's values such as emotions to do.

In addition, the product-specific value item evaluation result DB includes, for example, FIG.
(B) It is configured as shown.

FIG. 3A shows the information recommendation system users U1, U2, and U3 indicated by the identifiers V1 to V7 in FIG. 2 for the product represented by the identifier P1 in the product DB shown in FIG. It indicates the degree of importance of the value items listed. FIG. 3B shows the information recommendation system users U1 and U for the product represented by the identifier P2 in the product DB shown in FIG.
2, U3 indicates the degree to which the value items indicated by identifiers V1 to V7 are regarded as important. In other words, the product-specific value item evaluation result DB describes how multiple products / services held in the product DB are evaluated based on the values of multiple users based on consumption value theory. Is stored.

In the information recommendation system of the present invention, as shown in FIG. 10, the information acquiring means of the computer 1 acquires information from the information recommendation system user U1 or the like (step 101), and Is determined (step 10).
2) If the search condition is only for searching for a desired product from the product DB, a product meeting the search condition is selected and extracted from the product DB (step 103).
The selected and extracted merchandise is output from the information output means (step 105).

On the other hand, if the search condition determined in step 102 uses the value item DB 12, the information recommendation system user U1
First information (evaluation) indicating how much importance the information recommendation system user U1 or the like places on the plurality of “value items” V1 to V7 stored in the value item DB 12 (FIG. 2). The user is prompted to enter the information, and based on the first information obtained thereby, the information on the product corresponding to the first information is stored in the value-by-product evaluation result DB 13 (FIGS. 3A and 3B). ) And product DB
11 (FIG. 1) to perform an extraction process (step 104), and output the selected and extracted product from the information output means (step 105).

Here, the value item DB 11 (FIG. 2)
The first information (evaluation) indicating the degree of importance of the information recommendation system user U1 or the like with respect to the plurality of “value items” V1 to V7 stored in
For example, in FIG. 2, the information recommendation system user U1 and the like "value" or "not value" the value items represented by the identifiers V1, V2, V3, and the like.
(Recognition).

That is, the user who has accessed the information recommendation system of the present invention has a "value item"
The tendency of recognition (evaluation) of “emphasis” and “not emphasis” expressed for V1 to V7 is shown in FIG.
(B) It is determined which of the tendencies indicated by the information recommendation system users U1, U2, and U3 in the value item evaluation result DB for each product shown in the drawing is closer, and it is determined that the closer tendency is indicated. Product information, Product D
B (FIG. 1).

As a result, a user who has accessed the information recommendation system of the present invention can search for a product in which his or her own values (the way of thinking, feelings, etc., upon which to select a product) are more strongly reflected. And receive recommendations.

In the information recommendation system of the present invention, the processing section 5 further executes an information recommendation system user U by an information acquisition means as shown in FIG. 11 according to the computer program of the present invention.
After the information is obtained from the information recommendation system 1 (step 201), the information recommendation system provider stores the first information (evaluation) included in the information in the product DB 11 (FIG. 1). A plurality of “value items” V1 to V7 previously assigned to the products P1 to P10
Are aggregated and analyzed for each of (Step 202),
It is configured to perform processing for accumulating the information generated by the tallying and analysis in the product-specific value item evaluation result DB 13 (step 203).

By adopting such a configuration, the information recommendation system user U1 etc.
After accessing the information recommendation system of the present invention, for each of the products P1P to 10 stored in the product DB 11 (FIG. 1), each of the plurality of “value items” V1 to V7 of the value item DB 12 is “emphasized”. The information (evaluation) that expresses “does or does not value” is counted and analyzed, and the value item evaluation result DB13 for each product (FIG. 3)
(A), (b)).

Therefore, a value item evaluation result DB13 for each product
Are enriched each time the information recommendation system user accesses and provides the first information. Then, the user who has accessed the information recommendation system of the present invention selects his or her own values (products) by performing search and recommendation using the product-specific value item evaluation result DB 13 as described above. Products and products that more strongly reflect the way of thinking, emotions, etc. on which they are based, and receive recommendations.

In the information recommendation system of the present invention, the storage unit 4 further stores the degree of importance for each of the plurality of value items V1 to V7 of the value item DB 12 for each information recommendation system user U1 and the like. An area for a user-specific value item evaluation result DB 14 in which (evaluation) is accumulated is provided.

Further, the processing unit 5 further operates according to the computer program of the present invention as shown in FIG.
After the first information is obtained from the information recommendation system user U1 or the like by the information obtaining means (step 301), the first information included in the information is converted into a plurality of value items V1 in the value item DB12. To V7, the degree of importance of each information recommendation system user U1 and the like is counted and analyzed (step 302), and the plurality of value items V1 to V7 of the value item DB 12 as a result of the counting and analysis are respectively calculated. Is stored in the area for the user-specific value item evaluation result DB 14 of the storage unit 4 as the user-specific value item evaluation result DB 14 in the information recommendation system user U1 and the like. The processing is performed (step 303).

FIGS. 4A, 4B, and 4C show an example of the user-specific value item evaluation result DB. FIG.
(A) is an information recommendation system user U1
Has accessed, and for each product P1P-10 stored in the product DB (FIG. 1), a plurality of "value items" V
Aggregate and analyze the degree of importance of each of the plurality of value items V1 to V7 in the value item DB from the information indicating that each of items 1 to V7 is "important" or "not important". Then, the results of the aggregation and analysis are displayed. FIG. 4B shows the totaling and analyzing results in the case of the information recommendation system user U2, and FIG. 4C shows the totaling and analyzing results in the case of the information recommendation system user U3. As described above, the user-specific value item evaluation result DB is a database (DB) that holds what values each user has.

The process in the information recommendation system of the present invention described with reference to FIGS.
After the process of step 104 described in step 0, the program can be executed according to the above-described computer program of the present invention.

In the information recommendation system of the present invention, the storage unit 4 further stores the plurality of groups when the information recommendation system user U1 or the like is divided into a plurality of groups according to the first information. ,
A group DB storing information on the degree of importance (evaluation) of each of the plurality of "value items" V1 to V7 of the information recommendation user U1 and the like grouped into each of the plurality of groups.
15 and an area for a recommendation DB 16 including the following first group information, second group information, third group information, fourth group information, and fifth group information. I have.

The processing section 5 further acquires information from the information recommendation system user U1 or the like by the information acquiring means as shown in FIG. 13 according to the computer program of the present invention (step 401). ), According to the first information included in the information, the information recommendation system user who has provided the first information using the group DB 15, (Step 402), and according to the first information, the product-specific value item evaluation result DB1
3 and the user-specific value item evaluation result DB 14, the following first group information and second group information:
Generates information including third group information, fourth group information, and fifth group information, and stores these in the area for the recommendation DB.
It is configured to include a value item segmentation processing unit 9 that performs a process of retaining the value as B (step 402).

Here, the first group information is, when the information recommendation system user U1 or the like is divided into a plurality of groups according to the first information, the divided group and the first group. Associating each information recommendation system user who has provided information with which of the plurality of groups each information recommendation system user who has provided the first information is classified And is configured as shown in FIG. 5A, for example.

[0120] The second group information includes information of the information recommendation system users belonging to each group when the information recommendation system users are divided into a plurality of groups according to the first information. , And information on the average value of the degree of importance for each of the plurality of “value items”, for example, as shown in FIG.
It is configured as shown.

Further, the third group information is stored in the product DB in each group when the information recommendation system user is divided into a plurality of groups according to the first information. This is information on the average value of the degree of importance for each of the plurality of “value items” for each product, and is configured, for example, as shown in FIG. 5C.

Next, the fourth group information includes, for each of the plurality of “value items” when the information recommendation system users are divided into a plurality of groups according to the first information. Each information recommendation system user belonging to each group is information on the degree of importance given to each user. For example, FIG.
It is configured like the user-specific value item evaluation result DB shown in (b) and (c).

Further, the fifth group information includes, for each of the plurality of “value items” when the information recommendation system users are divided into a plurality of groups according to the first information, The degree of importance of each information recommendation system user belonging to the group is information stored for each product stored in the product DB, for example, as shown in FIGS. 3A and 3B.
Is constructed like a product-specific value item evaluation result DB shown in FIG.

By adopting the above-described system configuration, users having similar values are grouped by the above-described segmentation processing performed by the value item segmentation processing unit 9 so that users having the same values belong to the same group. You.

As described above, the recommendation DB 16 including the first group information, the second group information, the third group information, the fourth group information, and the fifth group information is stored in each group. The evaluation results of products or information performed by the system user of the present invention belonging to the present invention are extracted from the product-specific value item evaluation result DB 13 and the user-specific value item evaluation result DB 14, and are totaled for each group.

As a result, it is possible to receive a recommendation of a target product or information based on the recommendation DB 16 in which information on the evaluation of a person having the same values as the user is held.

If the value of each user is known, not only can the user directly search or recommend the target product or information, but also can search or recommend other products or information satisfying the value. . For example, thereby, it becomes possible to search for a product or receive a recommendation by specifying conditions such as “European travel package that suits me”.

[0128] From the viewpoint of the provider of the product or information to be searched or recommended, what kind of values a user who gives high (low) evaluation to the product or information has. Be clear. As a result, consultation such as product planning and sales strategy can be performed for the product or information provider. Or, if there is a difference between the user assumed by the product or information provider and the user who actually evaluates (finds high values), the target audience Alternatively, it is possible to indicate that a value item to be emphasized is changed, and that a new value item is added to provided products or information.

Further, in the information recommendation system of the present invention, the processing section 5 further performs the following processing according to the computer program of the present invention as shown in FIG.

The information recommendation system provider uses the information recommendation system by the information acquisition means for each of the plurality of “value items” previously assigned to each product stored in the product DB 11. After acquiring the information from the user (step 501), the recommendation DB 16 according to the second information included in the information and indicating the degree of importance of the information recommendation system user.
And the product DB 11 to extract information about the product corresponding to the second information (step 502).

Then, the extracted information is output from the information output means (step 503).

As described above, the processing shown in FIG.
It is determined whether the value of the user based on the value item included in the information from the user obtained by the information obtaining unit is close to the characteristic of which group stored in the recommendation DB, and the user belongs to the group. Which product or information satisfies the user's request is selected based on the value item evaluation result for the group's product or information, and output to the user.

Here, the second information is stored in a value item DB (FIG. 2) which is previously assigned to each of the products P1 to P10 stored in the product DB (FIG. 1). Information indicating how much importance the user of the information recommendation system regards a plurality of "value items" V1 to V7, and is represented by, for example, identifiers V1, V2, V3, etc. in FIG. With respect to the value items, the information recommendation system user has decided that “the degree of importance: 1 (not at all important)”, “the degree of importance: 2 (not very important)”, “ Degree: 3 (neither emphasized or not emphasized) "," Degree of importance: 4 (slightly emphasized) "," Degree of importance: 5 (very important) " Degree of emphasis, like Ranking, or level put to are those expressed.

Note that the recommendation DB 16 contains
By making a change to the above-described computer program for executing the processing by the information recommendation system of the present invention, a plurality of groups 1, 2, 3, 4,
5, the following information may be stored for each. Evaluation items and their values that characterize the group The users belonging to the group and the evaluation results of the value items by user The aggregation results of the evaluation results of the value items by user belonging to the group (average, deviation) Performed by the users belonging to the group Result of evaluation of value items by product, totaled (average, deviation) of evaluation results of value items by product belonging to the group

By using such information, a user who accesses the information recommendation system of the present invention can search for a product and make a recommendation in which the sense of values is more strongly and appropriately reflected. Can be provided.

An example of the above processing will be described with reference to FIG.
Product-specific value item evaluation result D by the processing described in FIG.
B and a user-specific value item evaluation result DB are constructed. Then, the value item segmentation processing is performed by the processing described with reference to FIG. 13, and the users are grouped based on the value item evaluation results. For example, assume that the following groups are obtained.

Group 1: Escape type value items B, D, and G are high Group 2: Type that wants to return to that day Value items A, D,
F and G are high Group 3: For the time being leisure type Value items A, E,
High F Group 4: High self-development type value items A and C

The evaluation results of the users belonging to each of these groups are totalized and analyzed. For example, how much the value items are expected for each group and how satisfied each individual product is is analyzed based on the deviation value. As a result, FIG. 16 (value item evaluation result for product A of group 1) and FIG. 17 (value item for product B of group 1) for three types of travel package products (product A, product B, and product C) for group 1 Evaluation result), FIG.
It is assumed that the result shown in (value item evaluation result for commodity C of group 1) is obtained.

Here, when recommending a travel package to a new user U5, what value the user U5 has is examined by the recommendation request receiving unit based on the value items. That is, with respect to the value items V1 to V7 assigned to the products (travel packages) in the product DB 11, the evaluation in the above five stages is provided as the second information.

As a result of analyzing the second information,
5 belongs to group 1 and when it is determined that the degree of expectation for the value item G is the highest, the product U is selected as a recommended product for the user U5 and recommended.

Alternatively, user U5 belongs to group 1,
When it is determined that the degree of expectation for the value items B and D is the highest, the product U is selected as a recommended product for the user U5 and recommended.

Next, the information recommendation system of the present invention is further configured as follows.

The storage unit 4 further stores a plurality of “value items” V1 to V10 previously assigned by the information recommendation system provider to each of the products P1 to P10 and the like stored in the product DB11. For each of the above, according to the second information, which is included in the information obtained from the information recommendation system user U1 or the like by the information obtaining means and indicates the degree of importance of the information recommendation system user. The information recommendation system which has provided the second information includes an area for a grouping DB 17 in which information on a grouping method for grouping the users is stored, and an area for a user unit value model DB 18.

The processing section 5 further performs the following processing according to the computer program of the present invention as shown in FIG.

First, the information recommendation system provider provides a plurality of "value items" V1 to V7 previously assigned to the products P1 to P10 stored in the product DB 11 (FIG. 1). After acquiring information from the information recommendation system user U1 or the like by the information acquisition means (step 601), the information indicating the degree of importance included in the information and recommended by the information recommendation system user. A process of forming a new group using the grouping DB 17 according to the second information so that the information recommendation system user who has provided the second information is located at the center of the group (step 602). I do.

Next, the value item evaluation result DB1 for each product
3. Using the user-specific value item evaluation result DB 14 to generate sixth group information, which is information on the information recommendation system users belonging to the new group formed as described above, The information recommendation system user belonging to the new group formed as described above generates the seventh group information related to the average value of the degree of importance for each of the plurality of “value items”, as described above. For each of the products stored in the product DB in the formed new group, the eighth group information on the average value of the degree of emphasis on each of the plurality of “value items” is generated, as described above. Each of the information recommendation system users belonging to the new group formed in Then, ninth group information relating to the degree of importance is generated, and each information recommendation system user belonging to the new group formed as described above places importance on each of the plurality of “value items”. The degree generates tenth group information summarized for each product stored in the product DB, and generates the sixth group information, the seventh group information, and the Information consisting of eight group information, ninth group information, and tenth group information,
It is stored in the area for the user unit value model DB of the storage unit 4 as the user unit value model DB 18 (step 6).
03).

[0147] Then, the information recommendation system provider checks each of the products P1 to P1 stored in the product DB11.
The information recommendation system included in the information acquired from the information recommendation system user by the information acquisition means for each of the plurality of "value items" V1 to V7, etc., which have been previously assigned to P10 and the like. User unit value model DB1 according to the second information indicating the degree of importance of the user
8 and the product DB 11 to extract information on the product corresponding to the second information (step 604).

The information about the product corresponding to the second information extracted as a result is output via the information output means (step 605).

In the above description, the sixth group information is, for example, configured as shown in FIG. 6A, and the seventh group information and the eighth group information are, for example, It is configured as shown in FIGS. 6 (b) and 6 (c). The ninth group information is
It manages the average of the value item evaluation results of the information recommendation system users belonging to the newly formed group, and has the same configuration as the user-specific value item evaluation result DB. Further, the tenth group information manages the average of the evaluation results of the value items for each product of the information recommendation system users belonging to the newly formed group, and the structure is composed of the value for each product. It is the same as the item evaluation result DB.

That is, in the recommendation of goods and information, not only the recommendation DB 16 segmented and generated in advance, but also dynamic segmentation according to the values of the user, the user unit value model D
B18 is generated, and a group having similar values centered on an arbitrary user is newly created to make a recommendation.

That is, according to the processing shown in FIG. 15, the value item evaluation result for the product or information of the group to which the user belongs is extracted from the user unit value model DB 18, and which product or information satisfies the user's request. You can choose to be satisfied and make recommendations.

According to this configuration, the information recommendation system provider provides a plurality of "value items" V1 previously assigned to the products P1 to P10 stored in the product DB 11 (FIG. 1). To V7, included in the information obtained from the information recommendation system user U1 or the like by the information obtaining means, the second information indicating the degree of importance of the information recommendation system user. Accordingly, a new group is formed using the grouping DB 17 so that the information recommendation system user who has provided the second information becomes the center of the group.

Therefore, when the classification (clustering) is performed, a user who accesses the system of the present invention for information recommendation is placed at the farthest position of the cluster, and the characteristics of the cluster are as follows. Even if the recommendation of the product based on the cluster is performed without representing the characteristics of the user, it is possible to prevent a situation in which the product desired by the user is not necessarily recommended.

In addition, according to the above-mentioned second information, the information recommendation system user who has provided the second information using the grouping DB 17 is renewed so as to be located at the center of the group. The process of forming a group includes, for example, “a grouping method for performing a factor analysis so that a factor contribution rate to a value item evaluation result of an information recommendation system user who has provided second information is maximized”, “ Information recommendation system that has provided the second information Grouping method that defines the discriminant function so that the distance to the user's value item evaluation result is minimized, "Using the information recommendation system that has provided the second information Grouping method that performs cluster analysis so that the distance to the value item evaluation result of the elderly is minimized An execution program, such as allowed to held in the grouping DB, can be performed by selecting one in this.

As described above, in the processing shown in FIG. 15, 1) the second information included in the information from the user acquired by the information acquiring means is analyzed, and the evaluation of the value item of the user is performed. Generate a result 2) Based on the evaluation result for the generated value item, perform any one of the following processes A, B, and C to create a new group so that the user is the center. ： Perform factor analysis so that the factor contribution rate to the user's value item evaluation result is maximized.
B: A discriminant function is defined so that the distance to the user's value item evaluation result is minimized. C: A cluster analysis is performed so that the distance to the user's value item evaluation result is minimized. Then, a user unit value model DB 18 is generated,
Create a new group with similar values, with a focus on arbitrary users, and make recommendations.

Further, the information recommendation system of the present invention may further include the following configuration in addition to the above configuration.

The storage unit 4 further stores a plurality of “value items” V1 to V7 previously assigned by the information recommendation system provider to the products P1 to P10 and the like stored in the product DB11. The information recommendation system provider has a concept dictionary DB19 in which a plurality of predetermined "concepts" are stored in advance in association with the plurality of "value items" V1 to V7. A word included in a sentence included in information obtained from the information recommendation system user by the information obtaining means, and a correspondence relationship between the word and the plurality of predetermined "concepts"; It has a word concept analysis DB 20 in which information is stored, and further has an area for a sentence-concept DB 21.

The processing section 5 further performs the following processing according to the above-described computer program, as shown in FIG.

First, free text input by so-called chatting or the like is obtained as information from a user of the information recommendation system by the information obtaining means (step 701). Next, the sentence is divided into words, and the relationship between the divided words and the plurality of “concepts” is analyzed using the word concept analysis DB 20 (step 702). Then, a sentence containing a related word corresponding to a plurality of “concepts” is associated with the related “concept”, and the sentence is stored in the sentence-concept DB area of the storage unit 4. -Store as the concept DB 21 (step 703). Next, a plurality of "value items" V1
The concept dictionary corresponding to the “concept” corresponding to V7 to V7 and the word including the related word corresponding to the “concept” are stored in the concept dictionary D.
The information is extracted using B19 and the sentence-concept DB 21 (step 704), and the information thus extracted is output from the information output means (step 705).

Each of a plurality of value items V1 to V7 in the value item DB shown in FIG.
(V1, V2, V4, V6), "There is a discovery" (V3,
A plurality of “concepts” such as V5 and V7) are determined and corresponded in advance by the information recommendation system provider.

Therefore, the information recommendation system of the present invention is realized by a server computer which can be connected to a network such as the Internet, and a person (user) who uses the information recommendation system of the present invention
When a free sentence is provided and input by chat or the like, the concept dictionary DB 19 and the word concept analysis DB 20 are used to use words included in the free sentence input in the chat (for example, “ Are associated with a plurality of value items V1 to V7 in the value item DB, and the association is stored in the sentence-concept DB 21 stored as data.

Then, the user includes words of the concept related to the value item that he or she considers important in the free text input by chat held in the information recommendation system of the present invention. The sentence-
It can be extracted and read from the concept DB 21. As a result, using the information recommendation system of the present invention, evaluations such as “emphasis” and “not emphasis” on a plurality of value items held in the value item DB are used as information. As described above, the user who provides and inputs the information to the information acquisition means can extract and read a sentence including a concept word related to the value item that the user considers important. While reading and confirming that the information retrieval and recommendation by the information recommendation system of is based on their own values, the free texts entered by others through chat are The information recommendation system of the present invention can be used.

The information recommendation system of the present invention described using the processing operation shown in FIG.
It can also be realized by the configuration described below.

In this case, the aforementioned word concept analysis DB 20
Instead of the second word concept analysis DB22 described below
And the second sentence-concept DB 23, and the second sentence-concept DB 23, the second sentence-concept DB 23,
The system is configured to perform the processing operations described below using the concept dictionary DB 19 described above.

The second word concept analysis DB 22 stores a plurality of “concept patterns” and registered “part-of-speech appearance patterns”.

Here, the “concept pattern” refers to a “character string with part-of-speech”, which is a free-form sentence corresponding to the “concept” stored in the concept dictionary DB 19, which is a “character string with part-of-speech”. ]. Note that this free description is associated with the “concept” as a sentence including a word that has a relationship corresponding to a plurality of “concepts” predetermined by the information recommendation system provider of the present invention. Can be stored in the text-concept DB 21.

The “part-of-speech appearance pattern” is the “conceptual pattern”, that is, the “part-of-speech-attached character string”, in which “part-of-speech with a part-of-speech other than the keyword” is the only part-of-speech ignoring words. Refers to information.

Here, the “keyword” is predetermined by the information recommendation system provider of the present invention. For example, “part of speech” or “part of speech” centered on a certain “part of speech” Are used as the “keyword”, and the sequence of “part of speech” ignoring the word related to “part of speech” is used as the “keyword”.

For example, when “noun”, which is an example of “part of speech”, is defined as a “keyword”, the concept dictionary D
Examples of a plurality of “concept patterns” stored in the second word concept analysis DB 22 corresponding to the “concept” that is “concept” stored in B19 and a registered “part of speech appearance pattern” Are as follows. (First example) "Concept" There is room for "Concept pattern" There is room for "concept pattern"
Shimoichi + Ru <live> Registered “part of speech appearance pattern” <contents> + <attached> / mood <name> + <attached> / <shitaichi> + <lively> (second example) “concept” There is room for “conceptual pattern” fresh <content> + na <attachment> / feel <name> + to <attachment> / are <shitaichi> + ru <live> Registered “part-of-speech appearance pattern” <content> + <Attached> / Mood <Name> + <Attached> / <Shimoichi> + <Live>

Here, <contents>, <appendices>, <names>, <lower ones>, <lives>, etc. are <contents>: adjectives, <appendixes>: adjuncts, <name>: nouns, <Shimoichi>: Utilize lower one, <Kyoo>:
Inflectional endings, <gei>: Geographic proper noun, <solid>: personal proper noun.

In the above description, when a sequence of “parts of speech” centered on a certain “part of speech” and ignoring words related to the “part of speech” is used as a “keyword”, for example, , "<Name> + <with> / <lower one>"
Is used as a “keyword”.

The original text of the “concept pattern” (for example,
"I can feel fresh." And "I can feel fresh."
Assume that the sentence contains the words "feeling" or "mood" that are related to the "concept" that is the predetermined "concept", and correspond to the concept "contains". Can be stored in the sentence-concept DB 21.

The processing unit 5 replaces the processing shown in FIG.
0, the following processing is performed.

The information obtaining means obtains, from the information recommendation system user, free text (sometimes referred to as “free description text” in the present specification and drawings) input by so-called chatting or the like as information. (Step 801).

Next, a free description sentence is a character string with a part of speech,
That is, it is divided into "character string with part of speech" and analysis "character string with part of speech" is obtained (step 802). For example, if a free description sentence "Taste the atmosphere of New York" is entered in a chat, it is divided into "Character string with part of speech"
An analysis such as "/ New York <Gizou> + <Appendix> / Yo <Content> + Na <Appendix> / Atmosphere <Name> + is <Appendix> / Taste <Gowa> + U <Kiteo>" Character string with part of speech ". The processing unit 5 that performs the processing operation of step 802 is, for example,
It can be called a second morphological analysis unit.

Next, in the analysis “character string with part-of-speech” obtained by the division performed in the second morphological analysis unit, the part-of-speech except for the keyword is changed to the part-of-speech state in which the word is ignored. An analysis “part of speech appearance pattern”, which is the “attached character string” information, is obtained (step 803). Here, the “keyword” is a “part of speech” which is determined in advance by the information recommendation system provider of the present invention,
A sequence of "parts of speech" centered on a certain "part of speech" predetermined by the information recommendation system provider of the present invention. I can put it. For example,
If “noun”, which is an example of “part of speech”, is defined as “keyword”, the analysis “character string with part of speech” obtained by the division performed by the second morphological analysis unit as described above is performed. , "/ <Geo> + <with> / <vol> + <with> / atmosphere <
The analysis “part-of-speech appearance pattern” of “name> + <attached> / <gowa> + <live >> is obtained. The processing unit 5 that performs the processing operation of step 803 can be called, for example, a part-of-speech appearance order analysis unit.

Next, the analysis “part-of-speech appearance pattern” obtained by the above-mentioned part-of-speech appearance order analysis unit and the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB 22
And searches for a registered “part of speech appearance pattern” that is similar to the analysis “part of speech appearance pattern” (step 804).

The search for the registered “part-of-speech appearance pattern” in step 804 is performed according to a method predetermined by the provider of the information recommendation system of the present invention. For example, a method of judging from the similarity of the “part of speech appearance pattern” (for example, a similarity indicating how similar the order in which the parts of speech appear around the keyword), a method of judging from the length of the character string, "A, a, c, d, e, a method of sorting and determining in the order of the Japanese syllabary, or a combination of one or more of these methods.

The processing unit 5 that performs the processing operation of step 804 can be called, for example, a “part of speech appearance pattern” comparison search unit.

As a result of the processing operation in step 804, there is a similar registered “part-of-speech appearance pattern” in the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB 22. The analysis "character string with part-of-speech" related to the analysis "part-of-speech appearance pattern" obtained by the above-mentioned part-of-speech appearance order analysis unit, and the free description sentence are the "concepts related to the registered" part-of-speech appearance pattern " The second sentence-concept DB 23 is stored in the second sentence-concept DB area of the storage unit 4 in association with the "concept" corresponding to the "pattern" (step 805).

Next, sentences relating to “concepts” corresponding to a plurality of “value items” V1 to V7 and “concept patterns” associated with the concepts are stored in the concept dictionary DB19.
And the second sentence-concept DB 23 (step 806), and the information thus extracted is output from the information output means (step 807).

According to the system configuration for realizing the processing operation shown in FIG. 19 described above, a user (user) using the information recommendation system of the present invention provides and inputs free text by chat or the like. Then, using the concept dictionary DB19 and the word concept analysis DB20, the words included in the free text input in the chat (for example, the words “there is room” and “there is discovery”) are used. Are associated with a plurality of value items V1 to V7 in the value item DB, and this association is held in the sentence-concept DB 21 held as data. Then, in the free text input by the chat held in the information recommendation system of the present invention, the user includes a text including a concept word related to the value item that the user emphasizes. From the sentence-concept DB 21 and can be read.

On the other hand, according to the system configuration for realizing the processing operation shown in FIG. 20, when the user provides and inputs a free description sentence by chat or the like, the concept dictionary DB19 and the second The free sentence input in the chat is analyzed into the analysis “part-of-speech appearance pattern” using the word concept analysis DB 22 of “. The concept pattern related to the“ part-of-speech appearance pattern ”based on the analysis“ part-of-speech appearance pattern ” Is associated with a plurality of value items V1 to V7 in the value item DB, and this association is held in the second sentence-concept DB 23 held as data. Then, in the free text input in the chat held in the information recommendation system of the present invention, the user can use the “concept pattern”, “part of speech appearance” The sentence analyzed in the "pattern" can be extracted and read from the second sentence-concept DB 21.

Therefore, in the case of the system configuration for realizing the processing operation shown in FIG. 19 described above, a free description sentence can be obtained only based on the correspondence between words and concepts predetermined by the information recommendation system provider of the present invention. In contrast to the above, according to the system configuration for realizing the processing operation shown in FIG. 20, according to the system configuration which realizes the processing operation shown in FIG. In a part-of-speech character string, a part-of-speech pattern other than the keyword (noun) is selected based on the correspondence between the part-of-speech appearance pattern in which only words and parts of the part-of-speech are ignored and the concept of the free-sentence sentence Therefore, even if there is a correspondence between the word and the concept that the information recommendation system provider of the present invention did not recognize, the user input the word. Of reason descriptive sentence "with part of speech string (that is, the concept pattern)", free description statement corresponding from the "part-of-speech appearance pattern" is selected, the user can read it. This allows the user to freely check whether or not the information retrieval and recommendation by the information recommendation system of the present invention is being carried out by reflecting his / her own values. The information recommendation system of the present invention can be used while reading and confirming the sentence more effectively.

Note that the second word concept analysis DB described above is used.
The contents of 22 can be automatically created and enriched by configuring the system of the present invention to execute the processing operations shown in FIG.

First, a free description sentence input by a so-called chatting or the like from a user of the information recommendation system is obtained as information by the information obtaining means (step 901).

Next, a free description sentence is a character string with a part of speech,
That is, it is divided into "character string with part of speech" and analysis "character string with part of speech" is obtained (step 902). The processing unit 5 that performs the processing operation of Step 902 is, for example, described in Step 8 described above.
02, it can be called a second morphological analysis unit.

Next, in the analysis “character string with part-of-speech” obtained by the division performed in the second morphological analysis part, the part-of-speech except for the keyword is changed to the state of only the part of speech ignoring the word. An analysis "part-of-speech appearance pattern" as "character string" information is obtained (step 903). The processing unit 5 that performs the processing operation of step 903 can be called, for example, a part-of-speech appearance order analysis unit, similarly to the case of step 803 described above.

Next, the analysis “part-of-speech appearance pattern” obtained by the above-mentioned part-of-speech appearance order analysis unit and the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB 22
And extracts a registered “part of speech appearance pattern” that is close to the analysis “part of speech appearance pattern” (step 90).
4).

The extraction of the registered “part-of-speech appearance pattern” in step 904 is performed in the same manner as in step 804, according to a method predetermined by the information recommendation system provider of the present invention. For example, a method of judging from the similarity of the “part of speech appearance pattern” (for example, a similarity indicating how similar the order in which the parts of speech appear around the keyword), a method of judging from the length of the character string, "A, I, U, D,
(E) A method of sorting and determining in the order of the Japanese syllabary, or a combination of one or more of these methods.

The processing section 5 performing the processing operation of step 904 can be called, for example, a “part of speech appearance pattern” comparison / extraction section.

Next, the keywords in the registered “part-of-speech appearance pattern” extracted by the processing operation of the “part-of-speech appearance pattern” comparison / extraction unit in step 904 and the part-of-speech appearance order analysis unit in step 903 described above. The keywords in the analysis “part of speech appearance pattern” obtained by the processing operation are compared with each other, and when the two are different, the step 902 according to the analysis “part of speech appearance pattern” obtained by the above part of speech appearance order analysis unit is performed. In the second word concept analysis DB 22, the analysis “character string with part of speech” obtained by the processing operation in the second morphological analysis unit in “the concept pattern” related to the registered “part of speech appearance pattern” A new “concept pattern” associated with the associated “concept” is registered in the second word concept analysis DB 22 (step 9). 6). The processing unit 5 that performs the processing operation of step 906 can be called, for example, a concept pattern registration unit.

The processing operations of steps 901 to 906 are specifically performed, for example, as follows.

It is assumed that the free description sentence input in the chat is "Taste the atmosphere like New York." In the second morphological analysis unit, this is expressed as "/ New York <gei *> + <attachment> / you <capacity> + na <attachment> / atmosphere <name> +
Is divided into "character string with part of speech" of <attachment> / taste <gowa> + u <live >> (step 902). For the parts of speech other than "", the analysis is a part-of-speech appearance pattern in which only the parts of speech are ignored, "<ground solid> + <attached> / <volume> + <attached> / atmosphere <name> + <attached>> / <Gowa>
+ <Live >> (step 903).

Next, in the “part of speech appearance pattern” comparison and extraction unit, the above-mentioned analysis “part of speech appearance pattern” (
<Attachment> / <content> + <attachment> / atmosphere <name> + <attachment> / <gowa> + <live >>) and the registered “part of speech appearance” stored in the second word concept analysis DB 22 Pattern ". Here, according to the method of judging from the viewpoint of how similar the order of appearance of the parts of speech with the keyword at the center, the free description sentence "I can feel fresh" corresponding to the concept of "there is room". "<Part> + <attachment> / mood <name> + <attachment> / <
Shimoichi> + <live>. Therefore, this similar registered “part-of-speech appearance pattern” “<content> + <attached> / mood <name> + <attached> / <lower one> +
<Live> is extracted (step 904).

Next, the registered “part-of-speech appearance pattern” extracted by the “part-of-speech appearance pattern” comparison and extraction unit “<
(In this case, "mood") and the analysis obtained by the part-of-speech appearance order analysis unit.> + <Appended> / mood <name> + <applied> / <lower> + <live>"Part of speech appearance pattern"
(<Geo solid> + <Attached> / <Content> + <Attached> / Atmosphere <Name> +
<Attachment> / <Gowa> + <Alive >> and the keywords (in this case, “atmosphere”) are compared. If the two are different, the analysis “part of speech appearance pattern” obtained by the part of speech appearance order analysis unit at step 903 (<ground solid> + <attached> / <content>)
+ <Attachment> / atmosphere <name> + <attachment> / <gowa> + <live >> analysis “character string with part-of-speech” obtained by the division performed in the second morphological analysis unit in step 902 according to step 902 ("/" New York <gei> + "" / "you""content" + "<">
/ Atmosphere <name> + <attachment> / taste <go> + u <live >> to the registered “part of speech appearance pattern” (<content> + <Appendix> / Mood <Name> + <Appendix> / <Shimoichi> + <Live >>"Conceptualpattern" (Fresh <Volume> + <Appendix> / Mood <Name> + <
Registered in the second word concept analysis DB 22 as a new "concept pattern" corresponding to the "concept" (there is a space) with which "attached" / are <the lower one> + ru <live> Step 906).

As a result, it is included in the “concept pattern” stored as a “character string with part-of-speech (that is,“ concept pattern ”) that has a relationship corresponding to the“ concept ”predetermined by the recommendation system provider. Not a new "concept pattern", "/ New York <gei *> + <with> / like <content> + na <with> / atmosphere <
The name> + <attachment> / taste <gowa> + u <live> is newly stored in the second word concept analysis DB 22 in association with the concept of “there is room”. The second word concept analysis DB 22 is automatically created, and “concept patterns” and “keywords” not initially recognized by the recommendation system provider are accumulated in the second word concept analysis DB 22, It is becoming.

It should be noted that step 90 in the processing operation described above is performed.
Between Step 4 and Step 906, a system configuration in which Step 905 described below is executed may be adopted.
This is because the frequency of use of the “concept pattern” used in the past for the “concept pattern” stored in the second word concept analysis DB 22 is analyzed, and only when the use frequency of the “concept pattern” is larger than a predetermined number, the step 906 is executed. If the frequency of use of the "concept pattern" is less than the predetermined number, the process does not proceed to step 906.

That is, after step 904, the keyword in the registered “part-of-speech appearance pattern” extracted by the “part-of-speech appearance pattern” comparison extraction unit and the analysis “part-of-speech appearance pattern” obtained by the part-of-speech appearance order analysis unit Is compared with the keyword in “.”, And even if both are different, the “concept pattern” stored in the second word concept analysis DB 22 corresponding to the registered “part of speech When the frequency of use is lower than the predetermined threshold, the analysis obtained by the second morphological analysis unit related to the analysis "part-of-speech appearance pattern" obtained by the part-of-speech appearance order analysis unit The “character string with part of speech” is changed to the “concept” corresponding to the “concept pattern” related to the registered “part of speech appearance pattern” extracted in the second word concept analysis DB 22. As a new "concept pattern" to be attached communication, it is not conducted process of registering the second word concept analysis DB 22.

Thus, in step 905, the registered “part of speech appearance pattern” extracted in step 904
For example, a processing unit that analyzes the frequency at which the “concept pattern” stored in the second word concept analysis DB 22 is used in the past in correspondence with the above can be referred to as a concept pattern use frequency analysis unit, for example.

As a specific process, the concept dictionary DB 19
The “concept pattern” stored in the second word concept analysis DB 22 corresponding to the “concept” that is “concept” stored in the second word and the registered “part-of-speech appearance pattern” are as follows: It is assumed that "Concept" There is room for "Concept pattern" / Self <name> / Likeness <tail> + <appended> / Found <lower><appended> Registered "part-of-speech appearance pattern" Self <name> / <tail> + <Appended> / <Lower><Appended>

The analysis "part-of-speech appearance pattern" obtained by the part-of-speech appearance order analysis unit in step 903 is the registered "part-of-speech appearance pattern" (self <name> / <tail> + <attachment> / <lower one><Appendix>), the registered “part-of-speech appearance pattern” extracted by the “part-of-speech appearance pattern” comparison and extraction unit in step 904 is “self <name> / <tail> +
<Attachment> / <bottom one><attachment>", and if the keyword in the analysis" part-of-speech appearance pattern "is different from" self ", the process proceeds to step 906 if normal, and the part-of-speech appearance order analysis The analysis “character string with part-of-speech” obtained by the division performed by the second morphological analysis unit corresponding to the analysis “part-of-speech appearance pattern” obtained by the second part is the second word concept analysis DB2
2. The registered “part of speech appearance pattern” extracted in 2
"Concept pattern" (/ self <name> / likeness <tail>) corresponding to (self <name> / <tail> + <attachment> / <lower one><attachment>)
+ Is a second word concept analysis DB as a new "concept pattern" associated with a "concept" (with a margin) associated with <attached> / <found> found <attached>)
22 is to be registered.

However, as a result of the analysis by the concept pattern use frequency analysis unit in step 905, this registered “part of speech appearance pattern” (self <name> / <tail> + <prefix> / <low one><prefix>) (Concept pattern) pertaining to (/ self <name> / likeness <tail> + <appended> / found <appeared below><applied>)
If the usage frequency in the past is smaller than a predetermined threshold, this registration process is not performed.

With the system configuration in which the processing operation of step 905 exists, the second word concept analysis DB
In the process of accumulating and enriching “concept patterns” and “keywords” in the second 22, a second sentence-concept DB 23
In the process of enriching the information, "concept patterns" and "keywords" that are more frequently used are accumulated, so that the reliability of the information recommendation system of the present invention is increased for the user. In addition, thereby, the information recommendation system of the present invention is automatically enriched with contents that are easier for the user to use.

As described above, the preferred embodiments of the present invention have been described with reference to the accompanying drawings. However, the present invention is not limited to such embodiments, and various modifications may be made within the technical scope described in the claims. The form can be changed.

For example, as an example of a structure of a value item DB suitable for extracting a user's sense of values, an example of modeling based on the consumption value theory of JNSheth has been described.
Is not limited to this content, and various forms can be adopted as long as they are suitable for extracting the values of the user.

In the information recommendation system of the present invention, when consulting on products or information is performed, the processing can proceed as follows. Receive any of the following instructions via the information acquisition means: a: target product or information / target product or value item to be provided by information b: target product or target user with information The following data is extracted from the recommendation DB 16: c: What value the target product or information is evaluated by the user to be evaluated d: High evaluation of the target product or information (low) ) What values do the user groups have? E: What is the user's evaluation of the value items that they intend to provide with the target product or information? F: With the target product or information What are the characteristics of a user group that has a high reputation for the value item to be provided? G: Target product Rui the user that are targeted at information (user group) is output through the information output means the data extracted from or recommendation DB16 are evaluating how the item or information.

[0208]

According to the information recommendation system of the present invention and a computer program for causing a computer to execute required processing for realizing the information recommendation system, information on a product is searched for, and a recommendation (information on recommended product information is obtained). The user who intends to receive the proposal) has a sense of his / her values (for example, the idea that the user depends on when selecting a product, the expectation and feelings of the service / product selected by the user, etc.) The search can be performed with due consideration given to the above, and the above-mentioned values can be recommended.

Further, according to the information recommendation system of the present invention and a computer program for causing a computer to execute required processing for realizing the information recommendation system,
Even if clustering (classification) is performed in advance and information is recommended based on the result, a new group is formed so that the user is at the center of the group, and the newly formed group is formed. Since the recommendation is performed with reference to the characteristic value evaluation for the group, the user can receive the recommendation in which the values of the user are sufficiently considered.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an example of a case where a product DB is configured as a travel package.

FIG. 2 is a value item D composed of a plurality of value items given to each product of a travel package.
The figure showing an example of B.

FIG. 3A is a diagram illustrating an example of a configuration of one product of a travel package in a product-specific value item evaluation result DB. (B) The figure showing an example of a structure about another one product of the travel package in the product-specific value item evaluation result DB.

FIG. 4A illustrates an example of a configuration of a user U1 in a user-specific value item evaluation result DB. (B) User U2 in the user-specific value item evaluation result DB
The figure showing an example of a structure about. (C) User U3 in user-specific value item evaluation result DB
The figure showing an example of a structure about.

FIG. 5A illustrates an example of a configuration of first group information. (B) The figure showing an example of the structure of the second group information. (C) A diagram illustrating an example of a configuration of third group information.

FIG. 6A illustrates an example of a configuration of sixth group information. (B) The figure showing an example of the structure of the seventh group information. (C) The figure showing an example of the structure of the eighth group information.

FIG. 7 is a view for explaining an example of the configuration of an information recommendation system of the present invention.

FIG. 8 is a view for explaining an example of the configuration of an information recommendation system of the present invention which is used by being connected to a network.

FIG. 9 is a diagram illustrating an example of a configuration of a storage unit in the information recommendation system of the present invention.

FIG. 10 is a flowchart illustrating an example of information recommendation and product search by the information recommendation system of the present invention.

FIG. 11 is a flowchart illustrating an example of building a value item evaluation result DB for each product in the information recommendation system of the present invention.

FIG. 12 is a flowchart illustrating an example of building a user-specific value item evaluation result DB in the information recommendation system of the present invention.

FIG. 13 is a flowchart illustrating an example of a segmentation process in the information recommendation system of the present invention.

FIG. 14 is a flowchart illustrating an example of another information recommendation by the information recommendation system of the present invention.

FIG. 15 is a flowchart illustrating an example of still another information recommendation by the information recommendation system of the present invention.

16 is a diagram illustrating a value item evaluation result for one group of products A newly formed in the processing illustrated in FIG. 15;

FIG. 17 is a diagram showing a value item evaluation result for a product B of one group newly formed in the processing shown in FIG. 15;

18 is a diagram illustrating a value item evaluation result for a product C of one group newly formed in the processing illustrated in FIG. 15;

FIG. 19 is a flowchart illustrating a process of extracting a concept from a free description sentence and matching a value item.

FIG. 20 is a flowchart illustrating another process of extracting a concept from a free description sentence and matching a value item.

FIG. 21 is a flowchart illustrating a process of automatically creating a second word concept analysis DB.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Computer 2 Information acquisition means 3 Information output means 4 Storage part 5 Processing part 6 Bus 7 Network 8 Communication line 9 Segmentation processing part U1, U2, U3, U4 User 11 Product DB 12 Value item DB 13 Value item evaluation by product Result DB 14 User Value Item Evaluation Result DB 15 Group DB 16 Recommendation DB 17 Grouping DB 18 User Unit Value Model DB 19 Concept Dictionary DB 20 Word Concept Analysis DB 21 Text-Concept DB 22 Second Word Concept Analysis DB 23 Second sentence-Concept DB

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Kazuhiko Nishimura 4-62 Higashinakano, Nakano-ku, Tokyo F-Term (in reference) 5B075 NK02 NK32 PR08 QT06

Claims (20)

[Claims] 1. An information recommendation system comprising: an information acquisition unit, an information output unit, a storage unit, and a processing unit, wherein the storage unit stores a product DB in which information on a plurality of products is stored. And, the information recommendation system provider, the product D
B, a value item DB in which information on a plurality of value items provided in advance is stored for each product stored in B, and the information recommendation system provider provides the product D
The degree to which the information recommendation system user regards each of the plurality of value items given in advance to each of the products stored in the product B in advance in the product DB stored in the product DB is determined. The processing unit stores the value item evaluation result DB for each product, wherein the processing unit accumulates in the value item DB included in the information obtained from the information recommendation system user by the information obtaining means. Information on a product corresponding to the first information based on the first information indicating how much importance the user of the information recommendation system regards the plurality of value items that have been set. A process of extracting using the value item evaluation result DB and the product DB is performed, and the product corresponding to the extracted first information is performed. Broadcast is, information recommendation system, characterized in that output via the information output means. 2. The information processing system according to claim 1, wherein the information recommendation system provider stores the first information included in the information acquired from the information recommendation system user by the information acquisition means in the product DB. For each product that is
It is configured to perform a process of totalizing and analyzing each of the plurality of value items provided, and accumulating information generated by the totaling and analysis in the product-specific value item evaluation result DB. The information recommendation system according to claim 1, characterized in that: 3. The storage unit further stores a user-specific value item evaluation result DB storing a degree of importance for each information recommendation system user for each of the plurality of value items in the value item DB. The processing unit further includes a first information included in the information obtained from the information recommendation system user by the information obtaining means, and a plurality of value items of the value item DB. Of the information recommendation system for each user, collect and analyze the importance of each user, and store the information generated by the totaling and analysis in the area for the user-specific value item evaluation result DB in the user-specific value field. 3. The information recommendation according to claim 1, wherein the information recommendation is configured to perform a process of retaining the item evaluation result DB. Yonshisutemu. 4. The storage unit further includes: a plurality of groups when the information recommendation system user is divided into a plurality of groups according to the first information; and a group of the plurality of groups. The information recommendation system includes a group DB in which information on the degree of importance of each of the plurality of value items of the user of the information recommendation system is stored, and an area for a recommendation DB. Further, according to the first information included in the information obtained from the information recommendation system user by the information obtaining means, the information recommendation system that has provided the first information using the group DB. When a user is divided into any of a plurality of groups in the group DB, In accordance with the first information, the product-specific value item evaluation result D
B and the user-specific value item evaluation result DB, and when the information recommendation system users are divided into a plurality of groups according to the first information, the divided groups and the first The information recommendation system users who have provided the first information are associated with each of the information recommendation system users, and each of the information recommendation system users who have provided the first information is grouped into any of the plurality of groups. First group information that is information indicating whether the information recommendation system user belongs to each group when the information recommendation system user is divided into a plurality of groups according to the first information. ,
The second group information, which is information about the average value of the degree of importance for each of the plurality of value items, and, according to the first information, when the information recommendation system user is divided into a plurality of groups. The third group information, which is information on the average value of the degree of importance of each of the plurality of value items for each product stored in the product DB in each group, According to the information, when the information recommendation system users are divided into a plurality of groups, for each of the plurality of value items, information on the degree to which each information recommendation system user belonging to each group regards them as important. According to certain fourth group information and the first information, an information recommendation system user For each of the plurality of value items when divided into number groups, the degree of importance of each information recommendation system user belonging to each group is stored for each product stored in the product DB. And the fifth group information, which is the first group information, the second group information, the third group information, the fourth group information, and the fifth group information. 4. The information recommendation system according to claim 3, further comprising a value item segmentation processing unit that performs a process of storing information as a recommendation DB in a recommendation DB area of the storage unit. 5. The processing unit further comprises: an information recommendation system provider, wherein the product D
The information recommendation included in the information acquired from the information recommendation system user by the information acquisition means for each of the plurality of value items previously given to each product stored in B. According to the second information indicating how much importance the system user takes, the recommendation DB
And processing for extracting information about a product corresponding to the second information using the product DB, and information about a product corresponding to the extracted second information is transmitted through the information output unit. The information recommendation system according to claim 4, wherein the information is output. 6. The storage unit, further comprising: an information recommendation system provider that stores the product D
The information recommendation included in the information acquired from the information recommendation system user by the information acquisition means for each of the plurality of value items previously given to each product stored in B. According to the second information indicating how much importance the system user takes, information relating to a grouping method for grouping information recommendation system users who have provided the second information is accumulated. And an area for a user unit value model DB. The processing unit further uses the grouping DB to store the second information in accordance with the second information. Form a new group so that the provided information recommendation system users are located at the center of the group, Using the product-specific value item evaluation result DB and the user-specific value item evaluation result DB, a sixth group that is information on information recommendation system users belonging to the new group formed as described above. Generating information, and generating seventh group information regarding an average value of the degree of importance of each of the plurality of value items of the information recommendation system user belonging to the new group formed as described above. Generating eighth group information relating to an average value of each of the plurality of value items that is regarded as important for each of the products stored in the product DB in the new group formed as described above; Each information recommendation system user belonging to the new group formed as described above, For each of the items, ninth group information relating to the degree of importance is generated, and each information recommendation system user belonging to the new group formed as described above assigns importance to each of the plurality of value items. The degree to which the product D
Generating tenth group information summarized for each product stored in B, generating the sixth group information, the seventh group information, and the eighth group information as described above. , The information comprising the ninth group information and the tenth group information is stored in the area for the user unit value model DB of the storage unit as the user unit value model DB, and according to the second information, And the user unit value model D
B, and information about a product corresponding to the second information is extracted using the product DB, and information about the product corresponding to the extracted second information is output via the information output unit. The information recommendation system according to claim 4 or 5, wherein the information recommendation system is performed. 7. The storage unit, further comprising: an information recommendation system provider that stores the product D
For each product stored in B, the information recommendation system provider sets a plurality of predetermined value items in advance in correspondence with the plurality of value items given and the plurality of value items. A concept dictionary DB in which concepts are stored; words included in a sentence included in information obtained from the information recommendation system user by the information obtaining means; And a word-concept analysis DB in which information on the correspondence relationship with a plurality of concepts is stored, and a region for a sentence-concept DB is provided. Divides the sentence included in the information obtained from the information recommendation system user into words, and uses the word concept analysis DB to Analyzing the relationship between the divided word and the plurality of “concepts”, associating a sentence containing a word having a relationship corresponding to the plurality of concepts with a concept having a corresponding relationship In addition, a text-concept DB is stored in the text-concept DB area of the storage unit as a text-concept DB, and a text including a concept corresponding to the plurality of value items and / or a word related to the concept is included. The information according to any one of claims 1 to 6, wherein the information is extracted using the concept dictionary DB and the sentence-concept DB, and a result of the extraction is output from the information output unit. Recommendation system. 8. The storage unit, further comprising an information recommendation system provider, wherein the product D
For each product stored in B, the information recommendation system provider sets a plurality of predetermined value items in advance in correspondence with the plurality of value items given and the plurality of value items. A concept dictionary DB in which concepts are stored, and sentences corresponding to “concepts” stored in the concept dictionary DB, and sentences stored in the sentence-concept DB are classified into “character strings with parts of speech”. A plurality of "concept patterns" and, in the "concept patterns", with respect to the parts of speech other than the keywords predetermined by the information recommendation system provider, the "character string with part of speech" information in which only the parts of speech ignoring the words are used. A second word concept analysis DB in which certain registered “part of speech appearance patterns” are stored, and a second sentence-concept DB area is provided. In addition, the processing unit further analyzes the sentence included in the information obtained from the information recommendation system user by the information obtaining means into a “character string with a part of speech” that is a character string with a part of speech. In the analysis "character string with part-of-speech", in the analysis "character string with part-of-speech", an analysis "part-of-speech appearance pattern", which is "part-of-speech character string" information in a state of only part-of-speech ignoring words for parts of speech other than the keyword, By comparing the analysis “part-of-speech appearance pattern” with the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB, the registered “part-of-speech pattern” similar to the analysis “part-of-speech appearance pattern” It is searched whether the "appearance pattern" exists, and the registered "part-of-speech appearance pattern" similar to the registered "part-of-speech appearance pattern" stored in the second word concept analysis DB The analysis “character string with part-of-speech” related to the analysis “part-of-speech appearance pattern” and the sentence corresponding to the “concept pattern” related to the registered “part-of-speech appearance pattern” The second sentence in the storage unit-
The concept dictionary stores, in the concept DB area, a second sentence-concept DB and a sentence relating to a concept corresponding to the plurality of value items and / or a “concept pattern” associated with the concept. The information recommendation according to any one of claims 1 to 6, wherein the information is extracted using a DB and the second sentence-concept DB, and a result of the extraction is output from the information output unit. Dation system. 9. The processing section further divides a sentence included in the information obtained from the information recommendation system user by the information obtaining means into a "character string with a part of speech" which is a character string with a part of speech. In the analysis "character string with part-of-speech", the analysis "character string with part-of-speech" information in which only the part-of-speech ignoring the word in the part-of-speech other than the keyword is analyzed "part-of-speech appearance pattern""
The analysis “part-of-speech appearance pattern” is compared with the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB, and the registered “part-of-speech pattern” approximated to the analysis “part-of-speech appearance pattern” The keyword in the extracted registered “part-of-speech appearance pattern” is compared with the keyword in the analysis “part-of-speech appearance pattern”, and when both are different, the analysis “part-of-speech appearance pattern” is performed. The analysis “character string with part-of-speech” according to the second word concept analysis D
A processing operation of registering in the second word concept analysis DB as a new “concept pattern” associated with the “concept” associated with the “concept pattern” related to the registered “part-of-speech appearance pattern” extracted in B 9. The information recommendation system according to claim 8, wherein 10. The processing unit further compares a keyword in the extracted registered “part-of-speech appearance pattern” with a keyword in the analysis “part-of-speech appearance pattern”. However, the frequency of use of the “concept pattern” stored in the second word concept analysis DB in the past corresponding to the registered “part of speech appearance pattern” is analyzed, and the use frequency is determined in advance. Only when the threshold value is exceeded, the analysis “character string with part-of-speech” related to the analysis “part-of-speech appearance pattern” is performed by the second word concept analysis D.
Registered “part of speech appearance pattern” extracted in B
10. The information recommendation according to claim 9, wherein a process of registering in the second word concept analysis DB as a new "concept pattern" associated with the "concept" corresponding to the "concept pattern" according to the above is performed. system. 11. A computer program that causes a computer including an information acquisition unit, an information output unit, a storage unit, and a processing unit to perform the following processing. The computer program is provided in the storage unit and relates to a plurality of products. Regarding a plurality of value items stored in a value item DB in which information on a plurality of value items preliminarily given to each product stored in a product DB in which information is stored, Based on the first information included in the information obtained by the information obtaining means, indicating whether the degree information recommendation system user attaches importance, the processing unit responds to the first information. Information about the product to be provided is provided in the storage unit, and the plurality of values assigned in advance to each product stored in the product DB. For each of the items, the degree of importance of the information recommendation system user is determined by using the product-specific value item evaluation result DB stored for each product stored in the product DB and the product DB. The information relating to the product corresponding to the extracted first information is output via the information output means. 12. The computer program, further comprising: a processing unit, wherein the first information included in the information obtained by the information obtaining means is added in advance to each product stored in the product DB. Tabulating and analyzing each of the plurality of value items,
12. The computer program according to claim 11, wherein the computer program causes the information generated by the analysis to be stored in the product-specific value item evaluation result DB. 13. The computer program, further comprising: in a processing unit, the first information included in the information acquired by the information acquiring means, for each of the plurality of value items in the value item DB, Of the importance of each person who provided the information of the information item, and analyzed the information.
For storing in the user-specific value item evaluation result DB area of the storage unit in which information regarding the degree of importance of each person who provided the first information for each of the plurality of value items is stored The computer program according to claim 11, wherein: 14. The computer program, further comprising: a processor configured to divide the information recommendation system users into a plurality of groups according to the first information provided in the storage unit. Using a group DB in which information on the degree of importance of each of the plurality of value items of the information recommendation user grouped into each of the plurality of groups is stored. According to the first information included in the information obtained by the means, the product purchaser who has provided the first information is divided into any one of a plurality of groups, the first When the information recommendation system user is divided into a plurality of groups according to the information, the divided group and the first The information recommendation system users who have provided information are associated with each of the information recommendation system users who have provided the first information. And generating first group information that is information indicating whether the product-specific value item evaluation result D according to the first information.
B and the user-specific value item evaluation result DB, using the information recommendation system belonging to each group when the information recommendation system user is divided into a plurality of groups according to the first information. Person of,
The second group information, which is information about the average value of the degree of importance for each of the plurality of value items, according to the first information, when the information recommendation system user is divided into a plurality of groups In each group, for each product stored in the product DB, third group information that is information on an average value of the degree of importance for each of the plurality of value items, Accordingly, when the information recommendation system users are divided into a plurality of groups, each of the plurality of value items is information on the degree of importance of each information recommendation system user belonging to each group. According to the fourth group information and the first information, the user of the information recommendation system is duplicated. For each of the plurality of value items when divided into groups, the degree of importance of each information recommendation system user belonging to each group is stored for each product stored in the product DB. 14. The method according to claim 13, further comprising the step of generating fifth group information, which is information stored therein, and performing a process of storing the information including the fifth group information in a recommendation DB area of the storage unit. Computer program. 15. The computer program, further comprising: a processing unit, for each of the plurality of value items, included in information obtained from the information recommendation system user by the information obtaining means, According to the second information indicating whether or not the user considers important, the recommendation DB and the product DB are used to perform a process of extracting information about a product corresponding to the second information, and the extracted 15. The computer program according to claim 14, wherein the computer-readable storage medium stores information on a product corresponding to the second information. 16. The computer program further comprises: a processing unit for each of a plurality of value items previously assigned by the information recommendation system provider to each product stored in the product DB; According to the second information, which is included in the information obtained from the information recommendation system user by the information obtaining means and indicates the degree of importance of the information recommendation system user, the second information The information recommendation system user who has provided the second information can use a grouping DB that stores information on the grouping method when grouping the information recommendation system users who have provided the information. Process to form a new group to be located at the center of Using the value item evaluation result DB and the user-specific value item evaluation result DB, the sixth group information that is the information of the information recommendation system users belonging to the new group formed as described above is obtained. The information recommendation system user belonging to the new group formed as described above, and generating seventh group information related to an average value of importance of each of the plurality of value items, For each of the products stored in the product DB in the new group formed as described above, eighth group information relating to an average value of the degree of importance for each of the plurality of value items is generated, and Each of the information recommendation system users belonging to the new group formed as described above, For this, a ninth group information relating to the degree of importance is generated, and the degree of importance of each information recommendation system user belonging to the new group formed as described above for each of the plurality of value items Is the product D
Generating tenth group information summarized for each product stored in B, generating the sixth group information, the seventh group information, and the eighth group information as described above. , The information comprising the ninth group information and the tenth group information is stored in the area for the user unit value model DB of the storage unit as the user unit value model DB, and according to the second information, And the user unit value model D
B, and using the product DB, extract information about the product corresponding to the second information, and output information about the product corresponding to the extracted second information via the information output unit. 16. The processing according to claim 14, wherein
Computer program as described. 17. The computer program further comprises: a processing unit that divides a sentence included in the information obtained from the information recommendation system user by the information obtaining means into words, and that the information recommendation system provider , The product D
For each product stored in B, the information recommendation system provider sets a plurality of predetermined value items in advance in correspondence with the plurality of value items given and the plurality of value items. By using a word concept analysis DB in which a plurality of concepts in a concept dictionary DB in which concepts are stored and information on correspondence relations with words included in the sentence are stored, the division is performed. Analyzing the relationship between the word and the plurality of concepts, and associating a sentence containing a word having a relationship corresponding to the plurality of concepts with the corresponding related concept, A sentence including a concept corresponding to the plurality of value items and / or a word having a relation corresponding to the concept is stored in the sentence-concept DB. Claim 1, characterized in that to perform the process of extracting with
A computer program according to any one of claims 1 to 16. 18. The computer program, further comprising: a processing unit that outputs a text included in the information obtained from the information recommendation system user by the information obtaining unit as a character string with a part of speech, The analysis is a "character string with part-of-speech" information, and in the analysis "character string with part-of-speech", the information is a "character string with part-of-speech" in which only words of the parts of speech other than the keyword are ignored. In addition to the "part of speech appearance pattern", the information recommendation system provider
For each product stored in B, the information recommendation system provider sets a plurality of predetermined value items in advance in correspondence with the plurality of value items given and the plurality of value items. A concept dictionary DB in which concepts are stored, and sentences corresponding to “concepts” stored in the concept dictionary DB, which are stored in the sentence-concept DB, are classified into “character strings with parts of speech”. A plurality of "concept patterns" and the "concept patterns"
Regarding the part of speech other than the keyword specified by the information recommendation system provider, the registered “part of speech appearance pattern”, which is the “string with part of speech” information in which only the part of speech is ignored, is stored. Using the second word concept analysis DB, the analysis “part of speech appearance pattern” is compared with the registered “part of speech appearance pattern” stored in the second word concept analysis DB, and the analysis “part of speech appearance A search is performed to determine whether a registered “speech appearance pattern” similar to “pattern” exists, and the registered “speech appearance pattern” stored in the second word concept analysis DB
The analysis "character string with part-of-speech" and the sentence related to the above-mentioned analysis "part-of-speech appearance pattern" in which the registered "part-of-speech appearance pattern" similar to In association with the “concept” corresponding to the “concept pattern” related to the “pattern”, the second sentence-concept DB is stored in the second sentence-concept DB area of the storage unit, and the plurality of values are stored. A process for extracting a sentence related to a concept corresponding to an item and / or a “concept pattern” associated with the concept using the concept dictionary DB and the second sentence-concept DB The computer program according to claim 11, wherein: 19. The computer program, further comprising: a processing unit for converting a text included in the information obtained from the information recommendation system user by the information obtaining means into a character string with a part of speech, The analysis is a "character string with part-of-speech" information, and in the analysis "character string with part-of-speech", the information is a "character string with part-of-speech" in which only words of the parts of speech other than the keyword are ignored The “part-of-speech appearance pattern” is compared with the analysis “part-of-speech appearance pattern” and the registered “part-of-speech appearance pattern” stored in the second word concept analysis DB. Registered "speech appearance pattern"
The keyword in the extracted registered “part-of-speech appearance pattern” is compared with the keyword in the analysis “part-of-speech appearance pattern”, and when both are different, the keyword related to the analysis “part-of-speech appearance pattern” Analysis "string with part of speech"
The second word concept analysis as a new “concept pattern” associated with the “concept” associated with the “concept pattern” related to the registered “part of speech appearance pattern” extracted in the second word concept analysis DB 2. A processing operation for registering in a DB.
8. The computer program according to 8. 20. The computer program, further comprising: comparing a keyword in the extracted registered “part-of-speech appearance pattern” with a keyword in the analysis “part-of-speech appearance pattern”; Even when the “concept pattern” stored in the second word concept analysis DB corresponding to the registered “part of speech appearance pattern” is used, the frequency of use of the “concept pattern” in the past is analyzed. Only when the value exceeds a predetermined threshold, the analysis “character string with part-of-speech” related to the analysis “part-of-speech appearance pattern” is performed using the registered “part-of-speech appearance pattern” extracted in the second word concept analysis DB. Is registered in the second word concept analysis DB as a new “concept pattern” associated with the “concept” corresponding to the “concept pattern” related to “. 20. The computer program according to claim 19, wherein: