JPH09153063A - Informatoin filtering device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a high-reliability relevance feedback by reflecting feedback information, which is obtained from each member, on a common profile. SOLUTION: The retrieval condition preliminarily designated for each of groups consisting of plural users 3 is held in a user information storage part 13, and feedback information which includes the goodness of fit decided by the user for text information presented by a text information output part 15 is inputted from a user information input part 11, and a user information analysis part 12 analyzes this feedback information to correct the retrieval condition for each group held in the user information storage part 13. Thus, high- realiability and efficient customization independent of individual emotions is realized.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information filtering apparatus for selecting from a huge number of text information that meets a user's request / interest and presenting it to the user.

[0002]

2. Description of the Related Art In recent years, with the spread of word processors and electronic computers, and the spread of electronic mail and electronic news via computer networks such as the Internet, the digitization of documents is accelerating. As the term "electronic publishing" implies, it is considered that information on newspapers, magazines and books will be generally provided electronically in the future. This allows
It is predicted that the amount of text information that will be available to individuals in real time will be enormous.

Along with this, there is an increasing demand for an information filtering system that selects from a vast amount of text information such as newspapers and magazines that meet the user's request and interest and provides it to the user at any time.

In the information filtering systems that have been realized so far, the user is made to evaluate the suitability of the presented text, and the result is fed back to a search condition called a profile for searching the text that is of interest to the user. By doing so, some have realized a relevance feedback function of increasing the suitability of the text for each user.

However, in the future, it is considered that the information filtering system will be utilized by a plurality of people who are interested in the same field, such as a research group having a specific theme of a research institute. The relevance feedback in the conventional information filtering system is only for responding to each user, and it is impossible to analyze the requests of a plurality of such users as a whole and provide the feedback. Also,
The textual relevance judgments made by individual users may not be consistent and reliable, so feedback does not necessarily improve text relevance.
It is desired to realize a more reliable relevance feedback function.

Further, in the conventional information filtering system, although a new search term can be added to the profile by relevance feedback, the search term which is no longer used in the text and is now out of date is automatically detected. It could not be deleted. Therefore, it is desired to realize an information filtering system that can follow the change of words and the change of topics.

[0007]

As described above, the relevance feedback in the conventional information filtering system is only for responding to individual users, and the demands of a plurality of such users are generally described. There was a problem that it was not possible to analyze and provide feedback.

Further, in the conventional information filtering system, although a new search term can be added to the profile by relevance feedback, the search term which is no longer used in the text and is out of date is automatically detected. There was a problem that it could not be deleted.

The present invention has been made in view of the above circumstances, and when a group of people who are interested in the same field share an information filtering system, the members of the group individually perform the information filtering system. Realizing reliable relevance feedback by enabling relevance feedback information to be reflected in a common profile, efficiently updating common profile and individual user's profile Furthermore, it is an object of the present invention to provide information filtering that realizes a search using only knowledge appropriate to the times by automatically deleting outdated search conditions and search terms in the profile.

[0010]

According to the present invention, in an information filtering apparatus for selecting desired text information from a plurality of text information and presenting it to a user, a search condition for each group formed by a plurality of users is set. Means for holding, means for selecting the text information that matches the search condition for each group held in this means, and means for presenting the text information selected by this means to the users forming the group A means for collecting feedback information which is a user's evaluation result for the text information presented by this means, and a means for correcting the search condition for each group based on the feedback information collected by this means. It is characterized by having.

In the present invention, feedback information from a plurality of users is comprehensively analyzed for each group, and the profile shared by the group is updated. In other words, the information of the relevance feedback that the members of the group individually performed is reflected in the common profile,
It is possible to customize information filtering system in a collaborative manner, and because it is based on the conformity judgment of multiple members, highly reliable relevance feedback that is not affected by the tendency of conformity judgment of individual users. Function can be realized.

Further, according to the present invention, in an information filtering device for selecting desired text information from a plurality of text information and presenting it to a user, a first condition for holding a search condition for each group constituted by a plurality of users is provided. Holding means, second holding means for holding the search condition for each user, means for selecting the text information that matches the search conditions held by the first and second holding means, and this means Means for presenting the text information selected by the above to the users constituting the group, means for collecting feedback information which is the evaluation result of the user with respect to the text information presented by this means, and means for collecting by this means A means for modifying the search conditions held in the first and second holding means based on the feedback information. And it features.

According to the present invention, the feedback information of each user is divided into the common profile updating information and the profile updating information for each user for feedback, so that a reliable relevance feedback function as a group is provided. While realizing it, information common to members is used to update the common profile as much as possible,
By using the information unique to other members to update the profile for each member, more suitable information filtering can be realized for each member, and the system throughput for feedback and information filtering The storage capacity for can be greatly saved.

Further, according to the present invention, in an information filtering device for selecting desired text information from a plurality of text information and presenting it to the user, means for holding search conditions, and the search conditions held by this means. Means for selecting the text information that matches with, means for presenting the text information selected by this means to the user, and time information for the search condition or search word that matches the text information presented by this means. It is characterized in that it is provided with means for adding and means for correcting the search condition based on the time information added by this means.

In the present invention, the time information is added to the search condition or the search word suitable for the search, and the user profile is corrected using this time information.
As a result, for example, by automatically deleting a search condition or a search term that has not been matched for a long period from the profile, it is possible to always perform a search using highly compatible knowledge.

Further, according to the present invention, in an information filtering apparatus for selecting desired text information from a plurality of text information and presenting it to the user, a means for holding a search condition and the search condition held by this means. Means for selecting the text information that matches with, means for presenting the text information selected by this means to the user, and feedback information that is the evaluation result of the user for the text information presented by this means is collected. Means, a first correction means for correcting the search condition based on the feedback information collected by the means, and a search condition or a search word newly added at the time of correction by the first correction means. And means for adding time information to the above, and second correction means for correcting the search condition based on the time information added by this means. Characterized in that was.

In the present invention, time information is added to the newly added search condition or search word at the time of relevance feedback, and the user profile is corrected using this time information. Thus, for example, by automatically deleting a search condition or a search term that has been long after being registered in the profile from the profile, it is possible to always perform a search using the latest knowledge.

[0018]

Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) First, a usage pattern of the information filtering system of the first embodiment will be described with reference to FIG.

As shown in FIG. 1, in this embodiment, five users 3 (users A, B, C, and 5) who have a common interest.
D, E) share the information filtering system 1. Then, text information arrives at any time from the text information source 2 to the information filtering system 1. For example, if users A to E are in a group that is conducting research on semiconductors and register "we are interested in text information about semiconductors" in the information filtering system 1, this is converted into a search condition called a profile. It After that, the information filtering system 1 automatically extracts only the text information regarding the semiconductor from the new arrival information and presents this text information to the users A to E.

In this embodiment, the information filtering system 1 has one common profile 10 for one group of users A to E, and narrows down information by using this common profile 10 as a search condition. Therefore, the text information presented to the users A to E in the figure is the same.

(System Configuration) FIG. 2 shows the functional configuration of the information filtering system 1 of this embodiment. In the figure, solid arrows indicate the flow of data.

As shown, the information filtering system 1 includes a text information analysis unit 16, a text information storage unit 17, a text information search unit 14, and a text information output unit 1.
5, a user information input unit 11, a user information analysis unit 12, and a user information storage unit 13. Among these components, the text information analysis unit 12, the text information search unit 14, and the user information analysis unit 16 surrounded by chain lines can be realized by software executed by the central processing unit of the computer, and the text information storage Part 1
3 and the user information storage unit 17 can be realized by a main storage device or a hard disk device of a computer. Further, the text information output unit 15 includes a CRT display or the like for presenting text information to the user 3,
The user information input unit 11 includes a keyboard and a mouse for inputting topics and relevance feedback information in which the user 3 is interested.

FIG. 3 shows an example of the processing flow of the text information analysis section 16.

The text information analysis unit 16 first takes in text information from the text information source 2 (step A).
1). Here, the text information source 2 is an institution such as a newspaper company or a publisher that generates text information and provides it to an information filtering system, or a text information source such as an electronic mail system or a document search system that handles text information separately. Systems, and sites that open the text information to the public on the computer network.

The text information analysis unit 16 performs morphological analysis, syntactic analysis, semantic analysis, format analysis and the like on the input text information, and frequency information and position regarding text constituent elements such as words, phrases, sentences and paragraphs. Information, text subject, and 5W1H-like information are extracted (step A2).

Then, each text is expressed by the extracted information (step A3). Then, the information extracted from the text information is converted into a format that can be searched by the text information search unit 14 (step A4) and stored in the text information storage unit 17 (step A5). This corresponds to the indexing process in the normal information search.

FIG. 4 shows an example of text information expressed by the text information analysis unit 16.

This figure shows a case in which a newspaper article stating that "○○ company and △△ company will establish a semiconductor joint venture company in XX prefecture XX city on the 17th of this month" is received from XX newspaper company. It is an expression example of the text information obtained in.

FIG. 5 shows an example of the processing flow of the text information search unit 14.

The text information search unit 14 first retrieves a profile, which is a search condition expressing the user's interest, from the user information storage unit 13 (step B1). If a plurality of users 3 have registered in the information filtering system 1 that they are “interested in information about semiconductors”, the profiles 10 for these users 3 include:
For example, the search condition as shown in FIG. 6 is described.

In FIG. 6, search conditions for searching text information related to "semiconductor" are listed.
“(Condition 1)” determines whether or not the word “semiconductor” appears in the text, and “(condition 2)” includes the words “memory” and “semiconductor”. It is to check whether or not they co-occur in the text. These correspond to the search conditions of the conventional pool search using the keyword. “(Condition 3)” is to search for text in which the company name of a specific semiconductor company appears in the heading of the text, and is a search condition that uses information about the appearance position of the word and knowledge about the company name of the semiconductor company. ing. “(Condition 4)” is a search condition based on word frequency information. “(Condition 5)” is a search condition that uses a related word in the field of “semiconductor” such as “DRAM” or “flash memory”.

The text information search unit 14 uses the profile 10 as shown in FIG. 6 as a search condition, sets the text stored in the text information storage unit 17 as a search target, and
A text search is performed (step B2). Here, the text search corresponds to, for example, selecting text information satisfying the search condition as shown in FIG. 6, and specifically, sorting the text information satisfying the search condition and the text not satisfying the search condition, It means to rank the texts according to the degree of satisfaction. For example, in the latter case, of the top text ranked by search,
Several were selected for user presentation, and the text information output section 1
5 (step B3). As a concrete method of the text search, for example, the technology disclosed in the literature (“SMART information search system”, edited by Gerald Sandton, translated by Shinken Kyoritsu, Planning Center) may be adopted.

FIG. 7 shows an example of the processing flow of the text information output unit 15.

The text information output unit 15 receives the text information to be presented to the user 3 from the text information retrieval unit 14 (step C1) and presents it to the user 3 (step C2).

FIG. 8 shows an example of the processing flow of the user information input unit 11.

The user information input unit 11 receives user information from the user 3 (step D1) and passes it to the user information analysis unit 12 (step D2). There are the following two types of user information.

The first is information about what kind of text information the user 3 wants, which the user 3 specifies in advance for the information filtering system 1, such as "I want an article about semiconductors". . Here, this type of information will be referred to as initial setting information.

The second is relevance feedback information in which the user 3 judges the suitability of the text information presented by the system. This is to modify the profile 10 so that the article to be presented to the user 3 is more suitable for the request of the user 3, and specifically, for example, information in the form as shown in FIGS. 9 and 10. Can be considered.

FIG. 9 is an example of information obtained by performing a three-step evaluation of "necessary / slightly necessary / unnecessary" on each of the text information presented by the user 3.

In this example, useful words are extracted from the words included in "text 1" and "text 2" which are determined to be "necessary" and added to profile 10.
By performing processing such as the above, there is a possibility that more filtering results will be obtained from the next time.

Further, as a modification of FIG. 9, the compatibility may be determined not in text units but in text component units. For example, extracting the sentence or paragraph of the text presented to the user 3, "this part was useful".
It is possible to feed back such information to the system. Furthermore, in FIG. 9, three-stage evaluation is performed, but this may be expanded to allow numerical evaluation of suitability.

FIG. 10 shows an example of relevance feedback information in natural language given by the user 3 via a keyboard or the like. An article about "semiconductor manufacturing equipment" was presented above the filtering results, but if user 3 is not very interested in "semiconductor manufacturing equipment", by returning a request such as (A) in the figure to the system, From the next time onward, it is conceivable that the texts including the word "semiconductor manufacturing equipment" will be reduced in score. Further, in the case of the search method in which the degree of importance is given to the search word in the profile 10, a request such as "emphasis on DRAM rather than flash memory" is issued as shown in (B) in the figure. Therefore, it is possible to change the importance of the search word in the profile.

FIG. 11 shows an example of the processing flow of the user information analysis unit 12.

The user information analysis unit 12 determines whether or not the information is already stored in the user information storage unit 13 (steps E1 and E2), and performs two operations according to the determination result. When the user information storage unit 12 is empty (Y in step E2), the initial selection information analysis process is performed (step E3), and when it is not empty, the relevance feedback information analysis process is performed (step E4).

FIG. 12 shows an example of the flow of initial selection information analysis processing.

In the initial selection information analysis processing, the user information analysis unit 12 analyzes the initial selection information by referring to the analysis dictionary 101 etc. necessary for language analysis prepared in advance,
Identify the word or expression that represents the selected topic (step F
1). Next, in order to obtain knowledge about the search word related to the selected topic and its synonyms, the search word is determined by referring to the topic knowledge prepared in advance (step F
2). Then, the profile 10 as shown in FIG. 6 is described using the determined search word. The generation of such a common profile 10 may be automatically performed by the system, or may be semi-automatically performed by causing the user 3 to correct the generated profile 10. The generated common profile 10 is stored in the user information storage unit 13 (step F3).

FIG. 13 shows an example of the topic knowledge described above.

FIG. 13A shows an example of topic knowledge describing the relationship between search terms. For example, since it is described that "ROM" and "RAM" are subordinate concepts of "semiconductor memory", when describing the profile 10 for the topic "semiconductor memory", "ROM" and "RAM" are described. Can be used as a search term. 13B is an example of topic knowledge regarding synonym information. By utilizing such knowledge and registering not only "ROM" but also a synonym "read only memory" in the profile as a search term, it is possible to perform a search that is not overlooked.

FIG. 14 shows an example of the flow of relevance feedback information resolution processing.

In the relevance feedback information folding process, the user information analysis section 12 analyzes the relevance feedback information of each user 3 by referring to the analysis dictionary 101 etc. necessary for language analysis prepared in advance. (Step G1 to step G4). Next, the feedback information to be reflected in the profile 10 common to the user 3 is selected from these pieces of information (step G5). Then, the common profile is updated using the relevance feedback method as disclosed in, for example, the document (“SMART Information Retrieval System”, edited by Gerald Salton, translated by God Health, and Planning Center), and this is updated as user information storage The data is stored in the section 13 (step G6). When a new word is added when updating the profile, the topic knowledge 102 about the topic may be referred to in order to obtain information about the related word of the word.

The difference between the conventional relevance feedback processing and the relevance feedback information analysis processing in this embodiment is that the former only reflects the feedback information obtained from a single user in a single profile. The latter means that the information to be reflected in the common profile 10 is selected from the feedback information obtained from a plurality of users and the feedback is performed.

FIG. 15 shows an example of relevance feedback information obtained from a plurality of users.

In this example, three users 3 who are interested in the same topic share the information filtering system 1, and the respective matching determination results for the three text information commonly presented to them are " It is indicated by “○” or “x”. Since users A, B, and C are interested in the same topic, it is generally considered that the determination results of each user 3 are the same, such as the compatibility determination results for “text 1” and “text 3”. . However, there may be discrepancies like the determination result of “text 2” in the figure due to slight discrepancies in the interests of the individual users 3, differences in knowledge about the field, or mood or busyness at the time of determination. Can be considered. In this example, although user A and user B have determined that “text 2” is “useful”, user C has determined that it is “unnecessary”. In such a case, for example, it is considered that highly reliable feedback can be performed if feedback processing is performed assuming that "text 2" is useful in a majority vote. Further, as a modified example, if the information that the user A is the most reliable suitability judge among the users A, B, and C in FIG. 15 is given to the system in advance, the feedback information of the user A is emphasized. Relevance feedback is also possible.

FIG. 16 shows a modification of relevance feedback information obtained from a plurality of users.

Such feedback information can be obtained, for example, by extracting words or phrases frequently appearing in the text or a part of the text that each user 3 has determined to be "useful". In this example, the text specified by the user A or a part of the text is "64 mega DRA.
It means that the words "M" and "semiconductor joint venture" were frequently used. The same applies to user B and user C. While terms such as “semiconductor joint venture”, “memory patent”, and “semiconductor manufacturing equipment” are included only in the feedback information of one user 3,
The word "64M DRAM" is included in the feedback information for all users. In such a case, "64 mega DRAM" is a word contained in the text or a part of the text determined to be useful by all users, and is the most important information to be reflected in the common profile 10. it is conceivable that. Therefore, "64M DRAM"
Only in the common profile, or in Figure 1
Even when all the words shown in FIG. 6 are reflected in the common profile 10, if the weight of “64 mega DRAM” is made higher than that of the other words, the reliability based on the majority vote of the plurality of users 3 is increased. Can provide high relevance feedback. As a specific reflection method to the profile 10, for example, “64 mega DRAM” is added to “(Condition 5)” in FIG. 6 or “(Condition 1)” to “(Condition 5)” is added. To all "64 mega DRA
For example, "M" may be added.

Although the relevance feedback based on the majority vote of the plurality of users 3 has been described above, it is possible to determine the information to be reflected in the common profile 10 from the feedback information of the plurality of users 3 according to a policy other than this. . For example, for each text information, if there is at least one user 3 in the group who finds it useful, the text information may be used for feedback. In this case, in the example of FIG. 15, it can be determined that “text 1” and “text 2” are useful, as in the case of majority vote. It is considered that such a policy may be effective when the group as a whole desires to perform filtering as thoroughly as possible. Similarly, for each text, if there is even one user 3 who does not need it, the text may not be used for relevance feedback. In this case,
In the example of 5, only "text 1" is adopted as the information suitable for relevance feedback. Therefore, the user 3 can sequentially specify the above-mentioned policy,
The method of selecting information to be used for relevance feedback may be switched according to the designated policy.

(Second Embodiment) Next, a usage pattern of the information filtering system of the second embodiment will be described with reference to FIG.

The difference between the first embodiment and this embodiment is that the former has one profile 10 common to a plurality of users 3 and one filtering result using this is presented to all users 3. On the other hand, the latter has both the common profile 10 and the profile 18 for each user 3, and finally presents the information customized for each user 3 to each user. The common profile 10 in this embodiment reflects the information request common to all users 3, and the user profile 18
Indicates that the information request unique to the user 3 is reflected. For example, users A to E in the figure are groups conducting research on the same theme, and even if they first register “I am interested in text information on semiconductors” in the information filtering system 1, they will be updated with time and new texts. Individual user 3 as information is taken in
Requirements may change in details. In the present embodiment, in order to deal with such a case, the relevance feedback information obtained from the plurality of users 3 is distributed to the information reflected in the common profile 10 and the information reflected in the user profile 18.

The device configuration of this embodiment is the same as that of the first embodiment shown in FIG. The functions of the text information analysis unit 16 and the user information input unit 11 are also the same as those described in the first embodiment. Here, only the points different from the first embodiment will be described.

FIG. 18 shows an example of the processing flow of the text information search unit 14 in this embodiment.

The difference between the flow of processing of the text information search unit 14 of the first embodiment and this embodiment is that the latter searches with the common profile 10 and individual user profiles 18 fused as the search condition. , To obtain search results for each individual user 3. That is, user A
For this purpose, a search condition is created from the common profile 10 and the user profile 18 for user A (step H4), a search is performed according to this search condition (step H5), and the searched text information is converted into text information. It is passed to the output unit 15 (step H6). Similarly, for the user B, a search condition is created from the common profile 10 and the user profile 18 for the user B to perform a search, and the searched text information is displayed as the text information output unit 15.
Pass to. This is performed for all users 3.

FIG. 19 shows an example of the processing flow of the text information output unit 15 in this embodiment.

The difference between the processing flow of the text information output unit 15 of the first embodiment and this embodiment is that the latter is user 3
It is to output the filtering result for each.

The user information analysis unit 12 in this embodiment.
Only the relevance feedback information analysis process is different from that of the first embodiment in the flow of the process (1), which will be described below.

FIG. 20 shows an example of the flow of relevance feedback information analysis processing in this embodiment.

The difference between the flow of the relevance feedback information analysis processing of the first embodiment and this embodiment is that, in the latter case, information that reflects relevance feedback information obtained from a plurality of users 3 in the common profile 10, The point is that the information is reflected to each user profile 18 before being fed back. This is shown in FIG.
An example will be described in which the feedback information as shown in FIG.

In FIG. 16, three users 3 who are interested in "semiconductors" individually conduct relevance feedback and extract "64M DRAM" from the information.
And "semiconductor joint venture". In this example, the feedback information of the three people is commonly “64 mega D
Since the word "RAM" has appeared, this word is not a word that indicates the fine taste of each user 3, but rather a search condition of the original search for the topic "semiconductor" that is more in line with the times. It may be helpful information to correct. For example, in the case where only the word "16 mega DRAM" appeared in the conventional text, but "64 mega DRAM" was newly developed and this word has become commonly used in the text. is there.
In such a case, the word “64 mega DRAM” is newly registered in the common profile 10, that is, the original search condition for obtaining general text information regarding “semiconductor”. More generally, the words commonly appearing in the feedback information of many users 3 are used for updating the common profile 10. For example, among the words shown in FIG. 16, if the words appearing in the feedback information of two or more people are used for the feedback to the common profile 10 and the remaining words are used for the feedback to the individual user profiles 18, “ 64 mega DRA
Only "M" will be fed back to the common profile 10 (step J5).

After the feedback regarding the common profile 10 is performed, the feedback processing regarding the individual user profile 18 is performed (step J).
7). In the example shown in FIG. 16, the word “semiconductor joint venture” is obtained as feedback information unique to user A. Therefore, the user profile 1 of user A
Register “Semiconductor Joint Venture” in 8. Similarly, “memory patent” and “SRAM” are set in the user profile 18 of the user B, and the user profile 18 of the user C is set.
Register "semiconductor manufacturing equipment".

As described above, if the feedback to the common profile 10 and the feedback to the user profile 18 are separately performed, the update of the general search condition of the original, which is necessary as time passes, Individual user customization can be done in one framework. When presenting a filtering result customized for each user 3 to a plurality of users 3, the conventional system has only the profile 18 for each user 3 and updates them individually, but in the present embodiment Then, the common feedback information is reflected only in one common profile 10, and the user profile 1
Since only the difference from the common profile 10 needs to be described in 8, it is more efficient in terms of processing amount and storage capacity.

(Display of Search Results Separated for Common Profile 10 and User Profile 18) In the present embodiment, the common profile 10 and the user profile 18 coexist, but it is conceivable to present information regarding this to the user. .

FIG. 21 shows an example of the common profile 10 and the user profile 18 for two users 3.

The common profile 10 includes (condition 1),
Search conditions such as (Condition 2), (Condition 3), ... And search terms corresponding thereto are described. The user profile 18 for the user A has (Condition A1), (Condition A)
Search conditions such as 2) and (condition A3) and search terms corresponding thereto are described. The search for the user A is performed by using both of them together. It is assumed that three articles for user A are obtained. Of these, "article 1" is an article that meets (condition 1) in FIG. 21, and "article 2"
22 is an article that meets (Condition 2) and (Condition 3), and “Article 3” is an article that meets (Condition A1) and (Condition A3), the article as shown in FIG. A presentation method can be considered.

As shown in the figure, the search results are divided into "articles that match the search conditions of the common profile" and "articles that match the search conditions of your personal profile". "Article 2" is the former,
"Article 3" is presented in the latter. This allows
The user 3 can easily know whether the presented article is adapted to the common interests of the group or to the personal interests.

FIG. 23 is a modification of FIG. In this example, the contribution level of the common profile 10 and the contribution level of the user profile 18 are added to each article. For example, “article 1” is common profile 1
Information that 30 partial points are given because 3 of 0 search conditions are satisfied, and 70 partial points are given because 7 of search conditions of the user profile 18 are satisfied is shown. Is displayed. In this example, it can be seen that “article 1” is the one most suited to the personal interests, and “article 3” is the one most suited to the interests common to the groups.

Information about the ratio between the size of the common profile 10 and the size of the user profile 18 may be provided together with the search results as shown in FIGS. 22 and 23. Here, the size of the profile refers to a value such as the number of search words or search conditions in the profile or the sum of the weights of the search words. For example, a user who has "words of common profile: words of user profile" 50:20,
It can be seen that it has different search results from the other members of the group, which reflects a much more personal interest than the 50: 3 user.

(Third Embodiment) Next, a usage pattern of the information filtering system of the third embodiment will be described with reference to FIG.

This embodiment is the same as the first and second embodiments except that a plurality of users 3 need not necessarily exist.

FIG. 25 shows a device configuration in this embodiment.

The difference from the first and second embodiments is that a user information management unit 20 is provided between the user information analysis unit 12 and the user information storage unit 13. The functions of the text information analysis unit 16, the text information output unit 15, and the user information input unit 11 are the same as those in the first embodiment. The functions of the user information analysis unit 12 are those shown in FIGS. 11, 12 and 14 only in that the analyzed user information is passed to the user information management unit 20 instead of being directly stored in the user information storage unit 13. Different from Therefore, here
Only the functions of the text information search unit 14 and the user information management unit 20 different from those of the first embodiment will be described.

FIG. 26 shows an example of the processing flow of the text information search unit 14 in this embodiment.

Here, a case of a search method in which the similarity between the profile 19 and the text information is calculated and the search is performed by ranking this will be described as an example. The text information search unit 14 performs the following process on all newly arrived text information.

First, the degree of similarity between the profile 19 and the text information is calculated according to the usual search method (step K3). Next, the current time is added to the search conditions in the profile 19 that match the text information in the similarity calculation described above (step K4). Similarly, the current time is added to the search words in the profile 19 that match the text information in the similarity calculation described above. These times added to the search condition and the search word will be called the latest matching time in this embodiment. When the text information search unit 14 completes the above processing for all the text information (Y in step K6), the text information is ranked in the order of similarity and the result is passed to the text information output unit 15 (step K7).
The function of the text information search unit 14 in the present embodiment,
The difference from that in the first embodiment is only that the former can enter the current time in the search condition and the search word in the profile 19.

FIG. 27 shows an example of the search condition and the search word to which the latest matching time is added.

As specific search conditions and search terms,
For example, the one shown in FIG. 4 can be considered. Here, there are four search conditions, and (search word A) to (search word F) are designated for each condition. The latest matching time is given to each search condition and search word. For example,
If certain text information includes (search term A) and (search term B) and these search terms satisfy "search condition 1" at "12:25 on 6th", (search condition 1) ,
(Search word A) and (search word B) are displayed as "6:12 2
The latest matching time of "5 minutes" is added. Next, if another text information includes (search term A), and this satisfies "search condition 1" at "11:30 on the 9th",
For (Search condition 1) and (Search term B)
The latest matching time of "0 minutes" is added. As a result, the profile 19 becomes as shown in FIG. As mentioned above,
The latest matching time indicates when the search condition or search word was recently matched. That is, the fact that this time is old means that the search condition or search term has not been used recently, and the fact that this time is new means that the search condition or search term is still effectively used in the search. It means that it is being broken.

FIG. 28 shows an example of the processing flow of the user information management unit 20 in this embodiment.

The user information management section 20 firstly makes the profile 19 to which the latest matching time as shown in FIG. 27 is added.
Is received from the user information analysis unit 12. Next, the latest matching time added to each search condition and search word in the profile 19 is compared with the current time. Then, the search condition and the search word that do not match any text information for a certain period are deleted from the profile 19, and the profile thus modified is stored in the user information storage unit 13. For example, assume that the current time is “12:30 on the 26th”, and the user information management unit 20 deletes the search condition and the search term that do not match for 20 days or more. At this time, the profile 19 as shown in FIG.
Is passed to the user information management unit 20, it is not used for more than 20 days because the latest compatible time is "6:12:25".
(Search term A) in (Search condition 1) and (Search term C) in (Search condition 2). Therefore, if these are deleted from the profile 19, an updated profile 19 as shown in FIG. 29 is obtained. Alternatively, as a modification, when weighting is given to each search condition or search word in the profile 19, it is possible to reduce only the weight without deleting the above-mentioned old search word.

With the above processing, new search terms are added to the profile by relevance feedback, while old search conditions and search terms that are no longer useful for the search are deleted from the profile. Therefore, it is considered possible to keep the profile in line with the times.

(Fourth Embodiment) Next, a fourth embodiment of the present invention will be described.

The usage pattern and device configuration of the information filtering system 1 in this embodiment are the same as those in the third embodiment. The functions of the text information analysis unit 16, the text information search unit 14, the text information output unit 15, and the user information input unit 11 are the same as those in the first embodiment. Therefore, only the functions of the user information analysis unit 12 and the user information management unit 20 will be described here.

FIG. 30 shows an example of the processing flow of the user information analysis unit 12 in this embodiment.

This substantially corresponds to that shown in FIGS. 11, 12 and 14 in the first embodiment, but when the profile 19 is first generated or when the profile 19 is updated by relevance feedback. The difference from the first embodiment is that the current time is added to each search condition and each search word. That is, in the initial selection information analysis processing (Y in step M2), the time when they are generated is added to each search condition and each search word in the profile 19 (step M5), and similarly the relevance feedback information is obtained. In the analysis process (step M
2N), the time is added to the search condition and the search word newly added by the relevance feedback method (step M7). These times added to the search condition and the search word will be referred to as profile registration times here. For example, it is assumed that the time of (search condition 1) is “12:25 on 6th” in a certain profile 19 and only (search word A) is designated as the search word corresponding to this condition. And, "9th, 11:30
It is assumed that (Search term B) is newly added to (Search condition 1) by relevance feedback. At this time, (search term B) is “9th, 11:30 pm”
The profile registration time is added, and at the same time, the profile registration time of (search condition 1) is also “9th, 11:30 pm”.
Minutes ”and the new profile looks like FIG. 26. As described above, the profile registration time indicates when the search condition or search term is registered in the profile.

FIG. 31 shows an example of the processing flow of the user information management unit 20 in this embodiment.

The user information management unit 20 first receives the profile 19 with the profile registration time as shown in FIG. 27 from the user information analysis unit 20 (step N1). Next, the profile registration time added to each search condition and search word in the profile 19 is compared with the current time (step N2). Then, the search condition and the search term that have been set for a certain period after being registered in the profile 19 are deleted from the profile 19 (step N
3) The profile 19 thus modified is stored in the user information storage unit 13 (step N4). For example, it is assumed that the current time is “12:30 on the 26th”, and the user information management unit 20 deletes the search condition and the search term that are 20 days or more after the registration. At this time, if a profile 19 as shown in FIG. 27 is passed to the user information management unit 20, it will be 2 after the registration.
Only the (search word A) of (search condition 1) and the (search word C) of (search condition 2) whose latest matching time is "12:25 on 6th" are 0 days or more. Therefore, if these are deleted from the profile, an updated profile as shown in FIG. 29 is obtained. Further, as in the case of the third embodiment, as a modified example, when each search condition or search term in the profile 19 is weighted,
It is conceivable to reduce the weight of the old search condition and search term without deleting them.

The difference between the third embodiment and this embodiment is that the former adds the search condition and the latest time when the search word matches the text to the profile at the time of search, whereas the latter adds the search condition and The point is that the time when the search word is registered in the profile is added at the time of relevance feedback. Also in the present embodiment, as in the third embodiment, new search terms are added to the profile by relevance feedback, while old search conditions and search terms that are no longer useful for the search are deleted from the profile. Therefore, it is considered possible to keep the profile in line with the times.

The information filtering apparatus according to the present invention described in the first to fourth embodiments is not constructed only in a distributed network environment, but is constructed in a personal computer operating in a single environment. It is possible.

[0097]

As described above in detail, according to the present invention, the relevance feedback information individually performed by the members of the group is reflected in the common profile, so that the customization of the information filtering system can be performed cooperatively. Can be performed. Further, the reliability of the feedback information is enhanced because it is based on the suitability judgment of a plurality of members. Further, since the information common to the members is used for updating the common profile as much as possible and the information unique to the other members is used for updating the profile for each member, the processing amount and storage capacity of the system can be saved. In addition, since old search conditions and search words are automatically deleted from the profile, it is possible to always search using the latest knowledge.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing a usage pattern of an information filtering system according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a device configuration of an information filtering system according to first and second embodiments of the present invention.

FIG. 3 is a diagram showing an example of a processing flow of a text information analysis unit according to the first and second embodiments of the present invention.

FIG. 4 is a diagram showing an example of text information expressed by a text information analysis unit according to the first and second embodiments of the present invention.

FIG. 5 is a diagram showing an example of a processing flow of a text information search unit according to the first embodiment of the present invention.

FIG. 6 is a diagram showing an example of a common profile according to the first and second embodiments of the present invention.

FIG. 7 is a diagram showing an example of a processing flow of a text information output unit according to the first embodiment of the present invention.

FIG. 8 is a diagram showing an example of a processing flow of a user information input unit according to the first and second embodiments of the present invention.

FIG. 9 is a diagram showing an example of relevance feedback information according to the first and second embodiments of the present invention.

FIG. 10 is a diagram showing an example of relevance feedback information according to the first and second embodiments of the present invention.

FIG. 11 is a diagram showing an example of a processing flow of a user information analysis unit according to the first and second embodiments of the present invention.

FIG. 12 is a diagram showing an example of the flow of initial selection information analysis processing in the first and second embodiments of the present invention.

FIG. 13 is a diagram showing an example of topic knowledge in the first and second embodiments of the present invention.

FIG. 14 is a diagram showing an example of the flow of relevance feedback information analysis processing according to the first embodiment of the present invention.

FIG. 15 is a diagram showing an example of relevance feedback information obtained from a plurality of users according to the first embodiment of the present invention.

FIG. 16 is a diagram showing a modified example of relevance feedback information obtained from a plurality of users in the first and second embodiments of the present invention.

FIG. 17 is a conceptual diagram showing a usage pattern of the information filtering system according to the second embodiment of the present invention.

FIG. 18 is a diagram showing an example of a processing flow of a text information search unit according to the second embodiment of the present invention.

FIG. 19 is a diagram showing an example of a processing flow of a text information output unit according to the second embodiment of the present invention.

FIG. 20 is a diagram showing an example of a flow of relevance feedback information analysis processing according to the second embodiment of the present invention.

FIG. 21 is a diagram showing an example of a common profile and a user profile for two users according to the second embodiment of the present invention.

FIG. 22 is a diagram showing an example of an article presentation method according to the second embodiment of the present invention.

FIG. 23 is a diagram showing a modified example of the article presentation method according to the second embodiment of the present invention.

FIG. 24 is a conceptual diagram showing a usage pattern of an information filtering system according to third and fourth embodiments of the present invention.

FIG. 25 is a diagram showing a device configuration of an information filtering system according to third and fourth embodiments of the present invention.

FIG. 26 is a diagram showing an example of a processing flow of a text information search unit according to the third embodiment of the present invention.

FIG. 27 is a diagram showing an example of a search condition and a search word to which the latest matching time / profile registration time is added in the third and fourth embodiments of the invention.

FIG. 28 is a diagram showing an example of a processing flow of a user information management unit according to the third embodiment of the present invention.

FIG. 29 is a diagram showing an example of a profile updated based on the latest matching time according to the third embodiment of the present invention.

FIG. 30 is a diagram showing an example of a processing flow of a user information analysis unit according to the fourth embodiment of the present invention.

FIG. 31 is a diagram showing an example of a processing flow of a user information management unit according to the fourth embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Information filtering system, 2 ... Text information source, 3 ... User, 10 ... Shared profile, 11 ... User information input part, 12 ... User information analysis part, 13 ... User information storage part, 14 ... Text information search part, 15 ... text information output unit, 16 ... text information analysis unit, 17 ... text information storage unit, 18 ... user profile, 19 ... profile, 20 ... user information management unit, 101 ... analysis dictionary, 102 ... topic knowledge,

Claims (4)

[Claims] 1. An information filtering apparatus for selecting desired text information from a plurality of text information and presenting it to a user, and a means for holding a search condition for each group formed by a plurality of users, and a means for holding the search condition. Means for selecting the text information that matches the held search conditions for each group, means for presenting the text information selected by this means to the users that make up the group, and the means for presenting the text information presented by this means Information comprising means for collecting feedback information, which is a result of user's evaluation of text information, and means for correcting search conditions for each group based on the feedback information collected by this means Filtering device. 2. An information filtering device for selecting desired text information from a plurality of text information and presenting it to a user, the first holding means holding a search condition for each group formed by a plurality of users. A second holding unit that holds a search condition for each user, a unit that selects the text information that matches the search conditions held by the first and second holding units, and a unit that is selected by this unit. Means for presenting the text information to the users forming the group; means for collecting feedback information, which is the evaluation result of the user for the text information presented by this means; and the feedback information collected by this means. And a means for modifying the search conditions held in the first and second holding means. Information filtering device. 3. An information filtering apparatus for selecting desired text information from a plurality of text information and presenting it to a user, said means for holding a search condition, and said means for matching said search condition held by said means. Means for selecting text information, means for presenting the text information selected by this means to a user, and means for adding time information to a search condition or search term that matches the text information presented by this means An information filtering device comprising: means for correcting the search condition based on the time information added by this means. 4. An information filtering device for selecting desired text information from a plurality of text information and presenting it to a user, and means for holding a search condition, and said means for matching the search condition held by this means. A means for selecting text information, a means for presenting the text information selected by this means to a user, a means for collecting feedback information which is a user evaluation result for the text information presented by this means, First correction means for correcting the search condition based on the feedback information collected by the means, and time information for the search condition or search word newly added at the time of correction by the first correction means. It is characterized by comprising means for adding and a second modifying means for modifying the search condition based on the time information added by this means. Information filtering apparatus according to.