JP2000067067A - Distribution server and distribution system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a distribution server and a distribution system for providing contents corresponding to the attribute of a user, for allowing the user to easily find desired information, and for reducing the increase of the load of a network and a distribution server. SOLUTION: This is a distribution server 4 for distributing contents, and this distribution server 4 is provided with a content pool 1 for storing plural contents, distribution filter 3 for receiving a content retrieving request including a user attribute from a client 2, retrieving the content pool 1 based on the user attribute, generating a content list with contents fulfilling a prescribed retrieval condition, and distributing the content list to the client 2, and content distributing means 13 for receiving a content request based on the content list from the client 2, reading the pertinent content from the content pool 1, and distributing it to the client 2.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a distribution server for performing PUSH type information distribution, and more particularly, to a distribution server for distributing content according to the attribute of a client (information receiver).

[0002]

2. Description of the Related Art FIG. 15 is a diagram showing a conventional system configuration, and shows broadcast type PUSH distribution. In the figure, reference numeral 4 denotes a distribution server, which has a content pool 1 in which a plurality of contents are stored. Reference numeral 2 denotes a client (a computer such as a personal computer used by a user), which is connected to the distribution server 4 via a line. When the distribution server 4 provides content to a plurality of clients 2 in the system configured as described above, the same content is simultaneously distributed to the plurality of clients 2. Note that the content includes at least a document reproduced (executed) by the client 2, and may further include a content posted on a web browser homepage.

[0003]

In the above prior art, since the same content is provided to a plurality of clients 2, content that is not particularly interesting to the user may be provided. In this case, there is a problem that it is difficult for the user to find information that he or she wants to know.
Further, in the above case, since the contents are transmitted in vain, the load on the network, the distribution server 4 and the client 2 increases.

[0004] The present invention has been made to solve the above-described problems, and can provide contents that match the attributes of the user, thereby making it easy for the user to find desired information. It is an object of the present invention to provide a distribution server and a distribution system capable of reducing the load on the network and the distribution server while making it possible.

[0005]

Means for Solving the Problems (1) A distribution server according to one aspect of the present invention is a distribution server that distributes contents to a client via a line, and includes a contents pool that stores a plurality of contents. Receiving a content search request including a user attribute from a client, searching a content pool based on the user attribute, generating a content list with contents satisfying a predetermined search condition, and distributing the content list to the client It is provided with a filter and a content distribution means for reading out the corresponding content from the content pool and distributing the content to the client when there is a content request based on the content list from the client.

(2) A distribution server according to another aspect of the present invention is a distribution server that distributes contents to a client via a line, and stores a plurality of contents each having one or a plurality of content attributes. When a content search request including a user attribute from a client is received from a content pool, a content having at least one of the content attributes matching the user attribute is searched from the content pool, and the corresponding content is searched. Generate a content list with
It is provided with a distribution filter for distributing the content list, and a content distribution means for reading out the relevant content from the content pool and distributing it to the client when there is a content request based on the content list from the client.

(3) A distribution server according to another aspect of the present invention is a distribution server that distributes contents to a client via a line, and includes a contents pool storing a plurality of contents, and a user from a client. When a content search request including an attribute is received, the content pool searches for a content that includes a character string that matches the character string of the user attribute in the content of the content or content-specific information that can identify the content of the content. A distribution filter that generates a content list and distributes the content list, and a content distribution unit that reads out the relevant content from a content pool and distributes the content to the client when a content request based on the content list is received from the client. It is.

(4) A distribution server according to another aspect of the present invention is a distribution server for distributing contents to a client via a line, wherein one or a plurality of contents attributes and thresholds are set. When a content search request including a user attribute from a client is received from a plurality of stored content pools, a semantic distance between the content attribute and the user attribute is calculated for each content attribute, and a semantic distance for each content is calculated. Finding the sum of the distances, searching the content pool for the content for which the sum of the semantic distances satisfying the predetermined condition with respect to the threshold is obtained from the content pool, generating a content list with the corresponding content,
It is provided with a distribution filter for distributing the content list, and a content distribution means for reading out the relevant content from the content pool and distributing it to the client when there is a content request based on the content list from the client.

(5) A distribution server according to another aspect of the present invention is a distribution server for distributing contents to a client via a line, comprising: a contents pool storing a plurality of contents for which a threshold is set; , A content search request including a user attribute is received, a semantic distance between each of the user attribute and the noun extracted from the content of the content or information capable of identifying the content of the content is calculated for each user attribute. Finding the sum of the semantic distances for each content, searching the content pool for the content for which the sum of the semantic distances that satisfy the predetermined condition with respect to the threshold is generated from the content pool, and generating a content list with the corresponding contents, A delivery filter that delivers the content list and content based on the content list from the client If there is determined, in which a content distributing means for distributing to the client reads the content of the relevant from the content pool. The predetermined condition in the above (4) and (5) is, for example, “exceeding the threshold” when the meaningful distance is given a large numerical value with respect to a meaningful distance. It is a condition or a condition of “not less than the threshold value”. Conversely, if the semantic distance is given a smaller numerical value than the meaningful distance, the condition is “less than the threshold” or “not more than the threshold”.

(6) A distribution server according to another aspect of the present invention is a distribution server that distributes content to a client via a line, and includes a content attribute indicating one or a plurality of pieces of position information, and a physical distance. When a content search request including a user attribute is received from a client, and a content search request including a user attribute is received from a client, a physical distance between the content attribute and the position information included in the user attribute is calculated. A content filter that searches the content pool for a physical distance that satisfies a predetermined condition with respect to the physical distance, generates a content list with the corresponding content, and distributes the content list with a distribution filter that distributes the content list; When there is a content request based on the content list of Is obtained by a content distribution means for distributing to the client reads from the pool.
The predetermined condition in the above (6) is, for example, a condition that is “within a physical distance” or a condition that is “within a physical distance”.

(7) In the distribution server according to another aspect of the present invention, the physical distance of (6) is obtained by referring to a physical distance database in which the physical distance between the position information is registered in advance. Things.

(8) A distribution server according to another aspect of the present invention is a distribution server that distributes contents to a client via a line, and stores a plurality of contents each having one or a plurality of content attributes. When a content search request including a user attribute from a client is received from a content pool and a client, a semantic distance between the content attribute and the user attribute is calculated for each content attribute. The average value of the semantic distance is defined as the semantic distance of the corresponding content, a content list is generated for each content including the semantic distance, and a content filter based on the content filter from the client and the content list from the client are used. When there is a content request, the corresponding content is read from the content pool It is that a content distribution unit to be delivered to the client Te.

(9) In the distribution server according to another aspect of the present invention, the semantic distance of (4), (5) or (8) is a meaning in which the semantic distance between nouns is registered in advance. It is obtained by referring to the database.

(10) In the distribution server according to one aspect of the present invention, in the distribution server according to any one of (1) to (9), ID information for each user and a user attribute are registered in association with each other. A database is further provided, and when a content search request including ID information is received instead of the user attribute, the corresponding user attribute is acquired from the database based on the ID information.

(11) A distribution system according to another aspect of the present invention transmits the distribution server according to any one of (1) to (10) and a content search request including a user attribute to the distribution server. Receiving a content list from the distribution server, displaying the content list, and when a desired content is selected based on the content list, transmitting a content request for requesting the selected content. is there.

The user attributes included in the content search request are those specified in advance by the distribution server.

The character string of the user attribute refers to the character string of the user attribute itself for a certain attribute item, and will be described with reference to an example of an embodiment described later. Point to.

[0018]

FIG. 1 is a diagram showing a system configuration of an embodiment of the present invention, FIG. 2 is a schematic diagram of FIG. 1, and FIG. 3 is an initial process until a client requests a content list from a distribution server. It is a figure for explaining the flow of.

The outline of the processing of this embodiment will be described with reference to FIG. As shown in FIG. 2, in the present embodiment, the distribution server 4 searches for a content suitable for the user based on the user attribute of the client 2 and the content pool 1 in which a plurality of contents are stored by a predetermined method.
A distribution filter 3 that performs so-called filtering that transmits only searched contents to the user is provided. In the present embodiment, the distribution filter 3 of the distribution server configured as described above is used to filter contents suitable for the user. By performing a search, generating a content list based on the searched content, and providing the content list and the content to the client 2 via a line, individual PUSH distribution by an individual attribute (user attribute) is realized. In the present embodiment, a distribution system is configured by the client 2 and the distribution server 4.

In FIG. 3, a client 2 is composed of a computer such as a personal computer used by a user, and has a display unit, an input unit such as a mouse, and a storage unit. The provided information providing program is installed. The information providing program provides content information to the user by using an image of a character or the like created by CG or the like and voice (text-to-speech synthesis). Individual attribute data (user attribute) is registered in the client 2. The user attributes indicate, for example, attributes of the user, such as the user's name, age, gender, and hobby, and are managed by the user and can be arbitrarily changed.

Here, the content will be described with a specific example. The content is a "character string to be spoken" composed of the content for making the character speak, and what kind of page is displayed on the Web browser. (Or the page itself) in which is described.

In FIG. 3, reference numeral 5 denotes a channel server, which is a device for instructing the position of the distribution server 4 to the client 2. The client 2, the channel server 5, and the distribution server 4 are connected via a line (not shown). It is connected.

Hereinafter, the flow of the initial processing until the client 2 requests the distribution server 4 for contents will be described with reference to FIG. The client 2 transmits a channel list request to the channel server 5, and when the channel server 5 receives the channel list request, transmits to the client 2 information indicating the positions of all the distribution servers 4 to the client 2 as a channel list. .
The client 2 displays the received channel list on a Web browser. Then, the client 2 transmits an initial scenario acquisition request to the desired distribution server 4 (FIG. 3).
Shows an example of transmission to distribution server 4 # 1 (channel 1) and distribution server 4 # 2 (channel 2).
This initial scenario acquisition request is for requesting an initial scenario for introducing a channel to a user or setting a client 2 relating to channel subscription prior to channel subscription. Information for setting the client 2 for the channel subscription includes:
For example, there is information for inquiring what user attributes are necessary to subscribe to the channel. Then, the distribution server 4 receiving the initial scenario request transmits an initial scenario response. If there are any user attributes required to subscribe to the channel other than the minimum required user attributes registered at the time of installation, a control statement instructing that additional user attributes be registered in the initial scenario The client 2 that has received the initial scenario displays, for example, a dialog or the like instructing to register the user attribute, and urges the user to register the user attribute.

The subsequent processing will be described with reference to FIGS. 1, 4 and 5. In FIG. 1, reference numeral 11 denotes an authentication server in which authentication data (for example, a name and a password for identifying each user) for authenticating a user is registered in advance, and an authentication filter 1 provided in the distribution filter 3.
In response to the inquiry about the individual attribute authentication from the server 2, the authentication result is returned. Reference numeral 13 denotes a content distribution unit that reads out the corresponding content from the content pool 1 based on a content request from the client 2 and transmits the content to the client 2. A database 14 stores user attributes and various data from the client 2.

FIG. 4 is a diagram showing the flow of processing between the client and the distribution server until the client obtains content that matches the user attributes. FIG. 5 is a diagram for explaining the operation of the client when the content list or the content is received.
The browser 22 is a character screen according to the information providing program.

Upon receiving the initial scenario response, the client 2 registers a user attribute based on the initial scenario,
Channel introduction etc. are performed. When subscribing to the channel, the client 2 transmits a channel initial information request (channel subscription declaration) to the channel (distribution server 4) to which the client desires to subscribe (S1). Distribution server 4
When receiving the channel initialization information request, the device transmits the channel initialization information to the client 2 (S2). The channel initialization information includes a user attribute list necessary for searching for content suitable for the user and generating a content list for each user.

Then, the client 2 transmits a content search request including the user attribute specified in the user attribute list to the channel to which the client desires to subscribe (S3).
The content search request includes authentication data (for example, a name and a password among the user attributes).
The distribution filter 3 is configured by the authentication filter 1
When the connection of the user is permitted as a result of the authentication, the content matching the user attribute is determined from the plurality of contents stored in the content pool 1 by a predetermined method. A search is performed to generate a content list including the corresponding content A, content C, and content D, and the content list is transmitted to the client 2 (S4). The processing for generating the content list by the predetermined method will be described later in detail.

Upon receiving the content list, the client 2 displays the character screen 22 to make the character speak the outline of the content list, and introduces the outline of the content list by images and sounds. When requesting appropriate content (for example, content A) based on this summary introduction, the distribution server 4 sends the content A
Is transmitted (S5). This content request is executed by, for example, performing an operation such as clicking the character screen 22 with a mouse at the time of introducing the outline of the content A by the character. Then, the distribution server 4 receiving the content request
The content distribution means 13 reads the content A from the content pool 1 and transmits the content A to the client 2 (S
6), the content A is executed by the client 2. That is, as shown in FIG. 5, the content is displayed on the Web browser 21 and the character is introduced to the content body. The case where the content B is read (S7, S8) is the same as the case where the content A is read, and therefore the description is omitted.

Although the case where the authentication server 11 and the authentication filter 12 are provided has been described above, they need not necessarily be provided. For example, in a case where contents are provided not to a specific user but to all users, the authentication server 11 is unnecessary.

Also, in the above description, a case has been described as an example where attribute data designated by the distribution server 4 is transmitted when a content search request is made. In this way, attribute data is added every time a content search request is made. Instead, attribute data can be registered in the distribution server 4 in advance. In this case, a database 14 for storing the attribute data is provided in the distribution server (see FIG. 1).

Hereinafter, the flow of processing in this case will be described with reference to the drawings. Since the flow is basically the same as that described above, different portions will be mainly described here. Here, the description will be made on the assumption that the authentication filter 12 and the authentication server 11 are not provided. FIG. 6 is a diagram showing a flow of processing between the distribution server and the client until the content is obtained when the attribute data is registered in the distribution server in advance.

The client 2 transmits a channel initial information request (channel subscription declaration) to the distribution server 4 (S
1) When receiving channel initialization information including a user attribute list (here, names, residences, and hobbies) from the distribution server 4 (S2), for example, "name = takeshi, residence = Tokyo, hobby" A user attribute registration request including the user attribute of “= baseball” is transmitted to the distribution server 4 (S1).
1). Upon receiving the user attribute registration request, distribution server 4 stores the attribute data in database 14. Thereafter, when the client 2 requests the distribution server 4 for the content list, the client 2 transmits a content search request including only the name to the distribution server 4 (S1).
2).

When the distribution server 4 receives the content search request, based on the name “takeshi” included in the content search request, the attribute data corresponding to “takeshi” from the database 14 (living place = Tokyo, hobby = baseball) Is acquired, and a plurality of contents stored in the content pool 1 are searched for contents that match the user attributes by a predetermined search method using the attribute data. Then, a content list (here, a list including content A, content C, and content D) is generated from the corresponding content, and the content list is transmitted to the client 2 (S4). The subsequent processing is the same as in FIG.
When an appropriate content (here, content A) is requested from the content list from the distribution server 4, a content request to that effect is transmitted to the distribution server 4 (S5). The content distribution means 13 reads the content A specified by the client 2 from the content pool 1 and transmits the content A (S
6), the content A is executed by the client 2. Although the information included in the content search request has been described using a name, the present invention is not limited to this. Information that can identify each user (for example, a user ID)
Is fine.

In the above description, it is described that the authentication filter 12 and the authentication server 11 are not provided. However, if these are provided, authentication data is required. Shall be transmitted.

Here, five specific examples of a processing method for generating a content list suitable for a user attribute will be described with reference to the drawings. First Example FIG. 7 is a flowchart showing the flow of a content list generation process (pattern matching). As shown in FIG. 7, each content is set with a content attribute which is information indicating what attribute should be distributed to the user. The content A is composed of information on traffic congestion in the city center, and is information for people living in Tokyo, and has a content attribute of “resident location = Tokyo”. The content B is composed of J league information, and is information for a person whose hobby is watching soccer or sports, and has a content attribute of "hobby = soccer, watching sports". The content C is composed of a professional baseball bulletin, and is information for a person whose hobby is watching baseball or watching sports.
A content attribute of "baseball, watching sports" is set. Content D is composed of event information at Tokyo Dome, where the residence is Tokyo or Kanagawa, and the hobby is information for baseball players, and the content attributes are "living location = Tokyo, Kanagawa" and "hobby = baseball" Is set.

The flow of the content list generation process (pattern matching) of the first example will be described below with reference to FIG. First, it is checked whether or not the content attribute matches the user attribute (S23). If they match, the process proceeds to the next content attribute (S24), and the process of step S23 is performed with the content attribute. This step S23 and S2
4 is repeated by the number of content attributes (S2
2) When the number of content attributes is repeated, this content (that is, content in which all content attributes match the user attribute) is added to the content list (S25), and the process proceeds to the next content (S26). It returns to the process of S22. However, if it is determined in step S23 that they do not match, the process exits from loop 2 and does not add the content to the content list.
The next content is transferred (S24), and the process returns to step S22. The processing between steps S22 and S26 is repeated by the number of contents (S21). When the number of contents is repeated, the contents list creation processing ends.

Hereinafter, a specific example will be described with reference to FIG.
In addition, it is assumed that the residence is set to Tokyo and the hobby is set to baseball in the user attributes. First, it is determined whether or not the residence of the content attribute in the content A matches the user attribute. Since both match in "Tokyo", the content A is added to the content list. The same determination is made for the content B, the content C, and the content D. Since the content B does not match any of the content attributes, the content B is not added to the content list.
Is added to the content list because the hobby matches "baseball", and the content D is added to the content list because the residence is "Tokyo" and the hobby is "baseball" and all the content attributes match.

In the above description, the content added to the content list has been described as having all content attributes matching the user attributes. However, not all content attributes but at least one content attribute matches the user attribute. The matching content may be used.

Second Example In the first example described above, content attributes are set for all contents. However, this setting process is complicated.
Also, an attribute that matches (characterly matches) the user's preference is not always set as a content attribute.

In the second example, the content of the content is searched based on the information of the user attribute without setting the content attribute particularly to each content, and when a matching character is found, the content is replaced by the user. Is determined to be information to be transmitted to the content list.

FIG. 8 is a flowchart showing the flow (pattern match) of the content list generation processing of the second example. As shown in FIG. 8, the content A is information on traffic congestion in central Tokyo,
Content B is composed of J league information.
Content C indicates a professional baseball bulletin, content D is composed of event information at Tokyo Dome, and content C and content D include “professional baseball”.

Hereinafter, the flow (pattern matching) of the content list generation process of the second example will be described with reference to FIG. First, it is checked whether or not a character that matches the user attribute character string exists in the content (S33). If not, the process proceeds to the next user attribute (S34), and the process proceeds to step S33 using the user attribute. Perform processing. The processes of steps S33 and S34 are repeated by the number of user attributes (S32). When the number of user attributes is repeated, the process proceeds to the next content (S35). However, if it is determined in step S33 that it exists, loop 2 is exited, and
Add the content to the content list (S3
5) Then, shift to the next content (S36), and step S
It returns to the process of 32. The processing of steps S32 to S36 is repeated by the number of contents (S31). When the number of contents is repeated, the contents list creation processing ends.

Hereinafter, a specific example will be described with reference to FIG.
It is assumed that hobby = baseball is set in the user attribute. First, it is checked whether or not the character of "baseball" exists in the content of the content A. Since the character does not exist, the same determination is made for the next content B. The same determination is made for the content C and the content D. Since the characters “baseball” exist in the content C and the content D, they are additionally registered in the content list.

Third Example In the first example, when the content attribute matches the user attribute, and in the second example, when a character matching the user attribute is found in the content of the content, the content is registered in the content list. I explained the case. Here, when the user attribute is, for example, hobby = sports, it is considered that the contents registered in the content list are the contents B and C, which are contents relating to sports (see the contents in FIGS. 7 and 8).
However, in the pattern matching of the first example, if there is no content whose content attribute is “sports”, no content is registered in the list. In the pattern matching of the second example, if “sports” that completely matches the user attribute is not included in the contents of the content B and the content C, neither is added to the list. Thus, for a user who does not have a content attribute or a user attribute that completely matches characters in the content, the content B and the content C are registered in the content list even though the content is related to sports. And will not be delivered.

In the third example, ambiguous semantic information search is performed, and contents having similar meanings even if they do not completely match characters are added to the contents list.

FIG. 9 is a flowchart showing the flow of the list creation processing (fuzzy search) of the third example. As shown in FIG. 9, for each content, a content attribute which is information indicating what attribute should be distributed to a user, and a threshold which is a reference as to whether or not to be added to the content list are set. ing.

Hereinafter, the flow of the list creation processing (fuzzy search) of the third example will be described with reference to FIG. First, the semantic distance between the user attribute and the content attribute is calculated (S
43) Then, the process proceeds to the next content attribute (S4).
4) Similarly, the process of step S43 is performed using the content attribute. The processes of steps S43 and S44 are repeated by the number of content attributes (S42). When the number of content attributes is repeated, the semantic distances acquired in step S43 are summed, and it is checked whether the sum exceeds a threshold (S45). If the sum exceeds the threshold, this content is added to the content list. (S46), the process proceeds to the next content (S47), and the process returns to step S32. If it does not exceed, the process proceeds to the next content without adding to the content list (S47), and returns to the process of step S42. The processing of steps S42 to S47 is repeated by the number of contents (S41). When the number of contents is repeated, the contents list creation processing ends.

It should be noted that the calculation of the semantic distance uses a technique such as a semantic information network which is an existing technique.

Kita Kenji, Satoshi Nakamura and Masaaki Nagata, "Spoken Language Processing: A Corpus-Based Approach", Morikita Publishing Co., Ltd., ISBN 4-627-82380-0, describe semantic distances.

Hereinafter, an example of the calculation method will be briefly described. FIG. 10 is a diagram illustrating an example of the semantic database. As shown in the figure, the semantic database stores a certain word (noun)
And a list of words related to the word, and the semantic distance between each word in the list and a certain word. In the figure, for example, "sky"
"Sky" in the list of words related to "sky"
"Moon,""sun,""star,""blue," etc.
The meaningful distance from the sky is 90, 20, 20, 2 respectively.
0 and 5 are registered. This semantic database is stored in the database 14, and the semantic distance is:
It is acquired by accessing the database 14 and referring to the semantic database. For example, the semantic distance between “sky” and “sky” is 90. The meaningful distance has a larger value as the meaning is closer.

Hereinafter, the flow of the content list creation process and the method of calculating the semantic distance will be described with reference to FIGS. 9 and 10. It is assumed that the residence is set to Hiroshima and the hobby is set to sports in the user attributes. First, in the content A, the semantic distance between the residence “Tokyo” of the content attribute and the user attribute “Hiroshima” is calculated. That is, the semantic distance 40 is acquired with reference to the semantic database 14. Then, it is checked whether or not the semantic distance exceeds the threshold value 40. In this case, the content A is not registered in the content list because it does not exceed the threshold value. Similarly, the semantic distance between the content B, the content C, and the content C is calculated, and the sum of the semantic distance of the content B is obtained as 70, and the sum of the semantic distance of the content C is obtained as 70. Will not be listed. Content D has a semantic distance between “Tokyo” and “Hiroshima” of 40 and a semantic distance between “baseball” and “sports” of 70. Will be added to the list.

Although the content to be added to the content list is a content in which the sum of semantic distances exceeds the threshold, the content may be, of course, a content having the threshold or more.
Further, in the above description, an example was described in which the closer the meaning, the closer the meaning, the larger the numerical value. However, conversely, the closer the meaning, the smaller the numerical value. Is content whose total semantic distance is less than the threshold or content that is less than or equal to the threshold.

Fourth Example In the third example described above, a case is described in which a fuzzy search is performed for the content attributes set for all the contents. First, processing for setting the content attributes for all the contents is described. Is complicated. In addition, since an attribute that is close to the user's preference and semantic distance is not always set as a content attribute, even if a fuzzy search is performed only on the content attribute, a content that matches the user attribute is searched. There was a part that could not be said.

In the fourth example, the contents of the content are searched based on the information of the user attribute without setting the content attribute to each content. The content is determined to be information to be delivered to the user, and is added to the content list.

FIG. 11 is a flowchart showing the flow (pattern match) of the list creation processing of the fourth example. FIG.
As shown in FIG. 1, the content of the fourth example is obtained by setting a threshold to the content of FIG.

Hereinafter, the flow of the list creation processing (pattern matching) of the fourth example will be described with reference to FIG. First, all nouns are extracted from the content (S5).
2). Then, the semantic distance between the user attribute and each of the extracted nouns is calculated (S54), the process proceeds to the next user attribute (S55), and the process of step S54 is similarly performed with the user attribute. Steps S54 and S55
Is repeated by the number of user attributes (S53).
Then, when the number of user attributes is repeated, the semantic distances acquired in step S54 are totaled, and it is checked whether the total exceeds a threshold (S56). If the total is exceeded, the content is added to the content list. (S57)
Shift to the next content (S58), and step S52
Return to the processing of. If it does not exceed, the content is not added to the content list, and the process proceeds to the next content (S58).
It returns to the process of step S52. The processing of steps S52 to S57 is repeated by the number of contents (S5
1) When the number of contents is repeated, the contents list creation processing ends.

Hereinafter, a specific example will be described. It is assumed that hobby = baseball is set as the user attribute, and the semantic distance is calculated using the semantic database of FIG. FIG. 12 is a diagram for explaining a specific example.
This example shows an example of the content D. Hereinafter, the content D will be described as an example. First, a noun is extracted from the content of the content D (see A in FIG. 12). Then, the semantic distance between the user attribute “baseball” and each of the extracted words (nouns) (see FIG. 12B) is calculated. FIG. 12C shows the semantic distance between “baseball” and each word, where the distance to “Tokyo Dome” is 15, the distance to “Yomiuri Giants” is 40, and the distance to other words is 0
Since the sum is 15 + 40 and exceeds the threshold value 50 of the content D, the content D is added to the content list.

Although the description has been made above assuming that a noun extracted from the contents of contents is ambiguously searched, for example, a title of contents (“contents A,“ city traffic information ”in the case of contents A)) or an outline may be used to identify contents of contents. .

Fifth Example The fuzzy search of the third and fourth examples is for searching for a word having a short meaningful distance between words (nouns). , Their physical distances can be very large. For example, both "Hiroshima" and "Tokyo" are nouns representing prefectures, and the actual physical distance is very large although the semantic distance is short. For this reason, in the case where the actual physical distance is significantly related to the distribution destination instead of the semantic distance as in the case of traffic jam information in central Tokyo, the above-mentioned fuzzy search has a portion that cannot be said to be an appropriate method. .

In the fifth example, in addition to a content attribute indicating position information such as a work place or a place of residence, a “distance” indicating how far from the place the user should be delivered is further provided. The search for the content to be added to the list is also performed in consideration of the typical distance.

FIG. 13 is a flowchart showing the flow of the content list creation processing (physical distance search) of the fifth example. As shown in FIG. 13, each content is set with a content attribute, which is information indicating what attribute should be distributed to the user, and a distance. Here, it is assumed that an attribute indicating position information such as a work place or a place of residence is set as the content attribute. Note that the “distance” indicates a distance from a location indicated by the position information. The content A is composed of information on traffic congestion in central Tokyo, and is information for a person whose residence or work location is within 50 km from Tokyo, and a content attribute of “residence = Tokyo or work location = Tokyo” and “distance = 50 k”.
m ”is set. The content B is composed of Yokohama beer hall information, and is information for a person whose work location is within 30 km from Yokohama, and a content attribute of “work location = Yokohama” and “distance = 30 km” are set.

Hereinafter, the flow of the list creation processing (physical distance search) of the fifth example will be described with reference to FIG. First,
The physical distance between the user attribute and the content attribute is calculated (S63). It is checked whether the physical distance calculated by this calculation is within the “distance” set for the content (S64), and if not, the process proceeds to the next content attribute (S65), and The processing of step S63 is similarly performed using the attribute. Steps S63 to S65
Is repeated by the number of content attributes (S62).
If it repeats, it will move to the next content (S66). However, if the physical distance exceeds the “distance” in step S64, the process exits from loop 2, adds the content to the content list (S67), shifts to the next content (S66), and proceeds to step S62. Return to processing. The process of steps S62 to S67 is repeated by the number of contents (S61).
The content list creation processing ends.

Hereinafter, a specific example will be described with reference to FIG. In addition, it is assumed that the residence is set to Tokyo and the work location is set to Urawa in the user attributes.
Within 0km and over 30km from Yokohama). First, in the content A, the physical distance between the residence “Saitama” of the content attribute and “Tokyo” of the user attribute is calculated. Since this physical distance exceeds 50 km, the work distance of the next content attribute is calculated. Similarly, the physical distance is calculated using “Tokyo” and the user attribute “Urawa”. Since the calculated physical distance is within 50 km, it is added to the content list. Then, for the next content B, the physical distance between the work location “Yokohama” of the content attribute and the user attribute “Urawa” of the user attribute is calculated. Since the physical distance exceeds 30 km, the content list creation processing ends without being added to the content list.

A physical distance database in which the distance between the position information is registered is stored in the database 14, and the physical distance is obtained by accessing the database 14 and referring to the physical distance database.

In the above description, the place of residence and the place of work have been described as examples of the content attribute relating to the position information. However, the present invention is not limited to this, and may be, for example, a postal code or a telephone number.

In the above description, the content to be added to the content list is a content whose physical distance is within “distance”, but may be a content whose distance is less than “distance”.

Sixth Example In the pattern matching and the fuzzy search described above, the number of contents transmitted to the client 2 becomes extremely small or too large, and it is not clear which content the user should select and execute. Problems could occur.

In the fifth example, the semantic distance between the user attribute and the content attribute is calculated, and the distance information is included in the content list so as to indicate the priority to the user.

FIG. 14 is a flowchart showing the flow of the content list creation process of the sixth example. As shown in the figure, each content is set with a content attribute, which is information indicating what attribute should be distributed to the user.

Hereinafter, the flow of the content list creation processing of the sixth example will be described with reference to FIG. First, the semantic distance between the user attribute and the content attribute is calculated (S7).
3) The process proceeds to the next content attribute (S74), and the process of step S73 is similarly performed using the content attribute.
The processes of steps S73 and S74 are repeated by the number of content attributes (S72). When the number of contents attributes is repeated, all contents including the distance information acquired in step S73 are added to the contents list. The processes of steps S72 to S76 are repeated by the number of contents (S71), and the content list creation process ends.

When there are a plurality of content attributes, an average value of the plurality of acquired semantic distances is used as the distance information.

Hereinafter, a specific example will be described with reference to FIG. In addition, it is assumed that the residence is set to Hiroshima and the hobby is sports, and the semantic distance is calculated using the semantic database of FIG. 10. First, in the content A, the semantic distance between “Tokyo” and “Hiroshima” is calculated and 40 is obtained.
, A semantic distance 70 between “soccer” and “sports” and a semantic distance 70 between “baseball” and “sports” in the content C are acquired. Then, content D
, The meaningful distance between "Tokyo" and "Hiroshima" is 40,
The semantic distance 70 between “baseball” and “sports” is acquired, and the average value 50 is acquired as distance information. Then, all contents including these semantic distances (distance information) are added to the content list, and the content list creation processing ends. D in FIG. 10 shows an example of a created content list.

In the above description, all the contents are added to the contents list. However, a predetermined number of contents may be selected from those having a higher numerical value of the distance information. Also,
In the case where there are a plurality of content attributes, it has been described that the average of a plurality of acquired semantic distances is included in the content list. May be.

According to the present embodiment, a content pool is searched based on a user attribute, a content list is generated based on the searched content, and the content list is distributed to a client. Content can be provided to the user. As a result, the user can easily find desired information, and as a result, it is possible to reduce an increase in load on the network and the distribution server 4.

As described in the first example, a search is made for a content in which all the content attributes set for the content match the user attribute, and the corresponding content is added to the content list. Content can be provided. This allows
The user can easily find desired information.

Also, as described in the second example, if the content attribute is not set for each content and the content of the content is searched based on the information of the user attribute and a matching character is found, Since the content is added to the content list, the complexity of the process of setting the content attribute can be reduced, and the content more suitable for the user can be provided. This allows the user to easily and reliably find desired information.

Further, as described in the third example, the semantic distance between the content attribute set for each content and the user attribute is calculated, and if the distance exceeds the threshold, the content is added to the content list. Therefore, even if the content attribute and the user attribute do not completely match with each other in character, it is possible to provide the user with content having a similar meaning.

Further, as described in the fourth example, the content is searched based on the information of the user attribute without setting the content attribute to each content, and a character with a short semantic distance is found. In that case, the content is added to the content list, so that the complexity of the process of setting the content attribute can be eliminated, and
Content suitable for the user can be provided. Thus, the user can easily find desired information.

As described in the fifth example, in addition to the content attribute indicating the location information such as the work location and the residence,
A "distance" indicating how far from the location the user should be delivered is further provided, and the physical distance between the content attribute and the user attribute is calculated, and this distance is within the "distance" set for the content. In this case, the content is added to the content list, so that the content can be provided in consideration of the physical distance.

Further, as described in the sixth example, the semantic distance between the user attribute and the content attribute is calculated, and the distance information is included in the content list. You can know the priority of the information you want.

As described above, according to the present invention, a content pool is searched based on a user attribute.
Since the content list is generated with the content satisfying the predetermined search condition and the content list is distributed to the client, the content suitable for the user can be provided to the user. This makes it possible to reduce the load on the network and the distribution server.

[Brief description of the drawings]

FIG. 1 is a diagram showing a system configuration according to an embodiment of the present invention.

FIG. 2 is a schematic diagram of FIG.

FIG. 3 is a diagram illustrating a flow of an initial process until a client requests a content list from a distribution server.

FIG. 4 is a diagram showing a flow of processing between a client and a server until content acquisition.

FIG. 5 is a diagram for explaining an operation of a client when receiving a content list or content.

FIG. 6 is a diagram showing a flow of processing between a distribution server and a client until content acquisition.

FIG. 7 is a flowchart showing a flow of a content list generation process (pattern match) of the first example.

FIG. 8 is a flowchart showing a flow (pattern match) of a content list generation process of a second example.

FIG. 9 is a flowchart showing the flow of a list creation process (fuzzy search) of the third example.

FIG. 10 is a diagram illustrating an example of a semantic database.

FIG. 11 is a flowchart illustrating a flow (pattern match) of a list creation process of a fourth example.

FIG. 12 is a diagram illustrating a specific example of the fourth example.

FIG. 13 is a flowchart illustrating a flow of a content list creation process (physical distance search) according to a fifth example.

FIG. 14 is a flowchart illustrating a flow of a content list creation process of a sixth example.

FIG. 15 is a diagram showing a conventional system configuration.

[Explanation of symbols]

 1 Content Pool 2 Client 3 Distribution Filter 4 Distribution Server 13 Content Distribution Means 14 Database

Claims (11)

[Claims] 1. A distribution server for distributing content to a client via a line, comprising: a content pool in which a plurality of contents are stored; and a content search request including a user attribute from the client. A content pool is searched based on the attribute, a content list is generated with content satisfying a predetermined search condition, and a distribution filter for distributing the content list to the client; and a content request from the client based on the content list is generated. A distribution server that reads out the corresponding content from the content pool and distributes the content to a client. 2. A distribution server for distributing content to a client via a line, comprising: a content pool storing a plurality of contents each having one or a plurality of content attributes; and a user attribute from the client. When a content search request is received, at least one of the content attributes is searched from the content pool for a content having a content attribute that matches the user attribute, a content list is generated with the corresponding content, and the content list is generated. A distribution filter for distributing, and content distribution means for reading out the relevant content from the content pool and distributing the content to the client when there is a content request based on the content list from the client. . 3. A distribution server for distributing content to a client via a line, comprising: a content pool in which a plurality of contents are stored; and a content search request including a user attribute from the client. Search the content pool for a content that includes a character string that matches the attribute character string in the content of the content or in content-specific information that can identify the content of the content, generate a content list with the corresponding content, and distribute the content list A distribution server, comprising: a distribution filter; and a content distribution unit that, when there is a content request based on the content list from the client, reads out the corresponding content from the content pool and distributes the content to the client. 4. A distribution server for distributing content to a client via a line, comprising: a content pool storing a plurality of contents each having one or more content attributes and thresholds; When a content search request including an attribute is received, a semantic distance between the content attribute and the user attribute is calculated for each content attribute, a sum of semantic distances for each content is obtained, and a predetermined value is determined with respect to the threshold. Searching the content from the content pool for which the sum of the semantic distances satisfying the condition is obtained, generating a content list with the corresponding content, and a distribution filter for distributing the content list, and based on the content list from the client If there is a request for the content, And a content distribution unit that reads out the content from the file and distributes the content to the client. 5. A distribution server for distributing content to a client via a line, comprising: a content pool in which a plurality of threshold-valued contents are stored; and a content search request including a user attribute from the client. Then, the semantic distance between each of the nouns extracted from the content of the content or information capable of identifying the content of the content and the user attribute is calculated for each user attribute, and the sum of the semantic distance for each content is obtained. A content filter that searches the content pool for a sum of semantic distances satisfying a predetermined condition with respect to the threshold, generates a content list with the corresponding content, and distributes the content list; When there is a content request based on the content list from Content distribution means for reading content from the content pool and distributing the content to the client. 6. A distribution server for distributing content to a client via a line, wherein a plurality of content items each of which has a predetermined physical attribute and a content attribute indicating one or a plurality of position information are stored. When a pool and a content search request including a user attribute from a client are received, a physical distance between the content attribute and the location information included in the user attribute is calculated, and a predetermined condition is set for the physical distance. A content filter that obtains a physical distance that satisfies the content pool is searched from the content pool, a content list is generated with the corresponding content, a distribution filter that distributes the content list, and a content request based on the content list from the client If there is, read the corresponding content from the content pool A distribution server comprising: a content distribution unit that distributes the content to a client. 7. The distribution server according to claim 6, wherein the physical distance is obtained by referring to a physical distance database in which a physical distance between position information is registered in advance. 8. A distribution server for distributing content to a client via a line, the content server including a plurality of content pools each storing one or more content attributes, and a user attribute from the client. When a content search request is received, the semantic distance between the content attribute and the user attribute is calculated for each content attribute. If the content has a plurality of content attributes, the average value of the semantic distance is calculated based on the semantic distance of the content. And generate a content list including the semantic distance for each content,
A distribution filter that distributes the content list; and a content distribution unit that reads out the corresponding content from the content pool and distributes the content to the client when there is a content request based on the content list from the client. Delivery server to do. 9. The method according to claim 4, wherein the semantic distance is obtained by referring to a semantic database in which semantic distances between nouns are registered in advance.
Or the distribution server according to claim 8. 10. A database in which ID information for each user and the user attribute are registered in association with each other, and when a content search request including the ID information is received instead of the user attribute, the ID information is received. The distribution server according to any one of claims 1 to 9, wherein the user attribute is acquired from the database on the basis of (1). 11. A distribution server according to claim 1, further comprising: transmitting a content search request including a user attribute to the distribution server; receiving a content list from the distribution server; A distribution system, comprising: a client that displays a list and, when desired content is selected based on the content list, transmits a content request that requests the selected content.