JP2001282838A - Contents recommendation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system for which information about age, gender, address, occupation, and hobby, etc., of each user is not indispensable, and which can recommend an appropriate content for individual user without requiring advance classification of contents. SOLUTION: A usage record device 24 records which content is used from user terminals 12-18. A content recommendation device 26, based on the usage record for each user, finds out contents used in combination, and defines the contents as a user-specific content group. In addition, based on the user-specific content groups, inclusive content groups in relation to every user are obtained. The content recommendation device 26, based on the user-specific content groups and the inclusive content groups, determines the recommended content for each user.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system capable of presenting appropriate recommended content for each user in a system capable of using a plurality of contents.

[0002]

2. Description of the Related Art Many websites provide many contents using the Internet. Since a large amount of content is overflowing, it is often difficult for users to find necessary content.

There is a method of uniformly recommending contents used by many users to each user as recommended contents. However, this method has a problem in that it is not possible to recommend content that matches each user.

[0004] Therefore, user information such as age, gender, address, occupation, and hobby of the user is registered in advance, the contents are classified by genre, and an appropriate classification is performed based on the user information. You can think of ways to recommend content.

However, in such a method, it is necessary to input a user's age, gender, address, occupation, hobby, and the like. For this reason, there is a possibility that use by users who value anonymity is not promoted. In addition, a large number of contents to be recommended had to be classified and managed, and the burden on the side providing the recommended information was heavy. In particular, it has been difficult to handle content that does not match the genre determined in advance.

The present invention solves the above-mentioned problems, and does not require information such as the age, gender, address, occupation, and hobby of the user, and does not require prior classification of the content. The purpose of the present invention is to provide a system capable of recommending appropriate content for each user.

[0007]

Means for Solving the Problems (1) A content recommendation system of the present invention provides a terminal device for browsing content and provides the content to the terminal device via a network in response to a request from the terminal device. A plurality of content providing means, a usage status recording means for recording content browsed by a user of the terminal device for each user, and a usage status for each user recorded in the usage status recording means for all users. Related acquisition means for acquiring the relationship between the contents based on the usage situation aggregated for each user, the usage situation of each user recorded in the usage situation recording means or information derived from the usage situation, and the relation acquisition means For each user,
Content recommendation means for determining recommended content and recommending the recommended content to the user of the terminal device.

[0008] As described above, the relation between contents is generated according to the browsing state of the contents, and the recommended contents are determined based on the use state of each user by using the relation information between the contents. . Therefore, appropriate content can be recommended for each user without requiring personal information of the user or pre-classification of the content. In addition, since the inter-contents related information also dynamically changes according to the browsing status, it is possible to automatically respond to changes such as addition of new contents.

(3) The content recommendation system according to the present invention is characterized in that the inter-content related information is grouping information in which each content is grouped.

(4) In the content recommendation system according to the present invention, the related acquisition means groups related contents for each user based on the usage status of each user recorded in the usage status recording means. Generates user-specific content group information, and generates comprehensive content group information as a grouping information based on the user-specific content group information. Yes, the content recommending means determines recommended content for each user based on user-specific content group information and comprehensive content group information.

Therefore, recommended content can be easily determined based on the difference between the content included in the user-specific content group and the content included in the comprehensive content group information.

(5) The content recommendation system of the present invention is characterized in that the inter-content related information is information indicating the degree of relevance between the contents.

(7) A content recommendation system according to the present invention includes a terminal device for browsing content, and a plurality of content providing means for providing content to the terminal device via a network in response to a request from the terminal device. A usage status recording unit that records the content browsed by the user of the terminal device for each user, and a user-related content based on the usage status of each user recorded in the usage status recording unit. Comprehensive content groups that generate user-specific content group information that groups content and that group related content for all users based on the user-specific content group information or the usage status of each user Related acquisition means for generating information, and user-specific content groups generated by the related information acquisition means Based on the flop information and the comprehensive contents group information, for each user, to determine the recommended content, and a content recommendation unit for recommending the user of the terminal device.

As described above, a content group for each user and a comprehensive content group are generated according to the browsing status of the content, and the recommended content is determined based on both. Therefore, appropriate content can be recommended for each user without requiring personal information of the user or pre-classification of the content. In addition, since the comprehensive content group also dynamically changes according to the browsing status, it can automatically respond to a change such as addition of new content.

(9) In the content recommendation system of the present invention, the use status recording means records the number of times each content is used for each user, and the content group generation means uses the number of times each content is used. Thus, user-specific content group information is obtained.

Therefore, recommended contents can be determined based on the number of times each content is used.

(10) In the content recommendation system according to the present invention, the use status recording means records use points in consideration of weighting based on the use type, and the content group generation means uses the use point of each content based on the use points. Thus, user-specific content group information is obtained.

Therefore, it is possible to determine the recommended content based on the usage point in consideration of the weight based on the usage type.

(11) In the content recommendation system according to the present invention, the usage status recording means records the usage time or day of each content for each user, and the content group generation means determines the usage time zone. Or, for each day of the week, the comprehensive content group information is obtained, and the content recommendation unit uses the comprehensive content group information for the time zone to which the time belongs or the day of the week according to the current time or the day of the week, for each user. It is characterized in that recommended contents are determined.

Therefore, it is possible to determine an appropriate recommended content for each time zone or day of the week.

(12) The content recommendation system of the present invention is characterized in that the same content is handled as different content for each time zone or day of the week.

Therefore, it is possible to recommend a content by designating a time zone or a day of the week.

(13) In the content recommendation system according to the present invention, the content group generating means includes: a user-specific content group and a comprehensive content group;
Grouping that considers the browsing order of each content,
The content recommending means determines recommended content for each user in consideration of the browsing order.

Therefore, it is possible to recommend contents including the browsing order.

(14) In the content recommending system according to the present invention, the content group generating means obtains comprehensive content group information for each attribute group based on the attribute of the user, and the content recommending means obtains the content group for each user. It is characterized in that recommended content is determined for each user based on the information and comprehensive content group information of the attribute group to which the user belongs.

As described above, by generating a comprehensive content group for each attribute group based on the user attribute, more preferable content can be recommended.

(15) In the content recommending system of the present invention, the content recommending means finds a user-specific content group constituted by a part of the content included in the comprehensive content group, and includes the content group in the user-specific content group. Content that is not included in the comprehensive content group
Characteristically determined as recommended content.

The content included in the comprehensive content group but not included in the user-specific content group is content that is viewed by many other people who are viewing similar content. The significance of selecting as recommended content is great. That is, appropriate content can be recommended.

(16) The content recommendation system according to the present invention is characterized in that the content group generating means recalculates the user-specific content group information and / or the comprehensive content group information at predetermined time intervals.

Therefore, it is possible to recommend an appropriate content in response to a change in the browsing situation, a change in the content, and the like.

(17) In the content recommendation system according to the present invention, the content group generating means determines the timing of the next recalculation based on the degree of change between the previous comprehensive content group information and the current comprehensive content group information. It is characterized by doing.

Therefore, the timing of the recalculation can be appropriately determined according to the degree of the change.

A "network" refers to a device for communicating with two or more devices, whether wireless or wired, and is not limited to an open device such as the Internet, but also an LA.
This is a concept that includes a closed one like N.

"Content related information" refers to information indicating the relationship between contents used for content recommendation. In the embodiment, the comprehensive content group information in FIG. 11 and the related diagram by the network in FIG. 18 correspond to this.

The term "per user" is a concept that includes not only the case of each user literally but also the case of each terminal device. That is, in a mobile phone or the like, the same terminal device is almost always used by the same user, and when each terminal device is used, the same terminal device is substantially used by each user.

"Connected" is a concept that includes not only a case where connection is always made by a dedicated line but also a case where connection is possible when necessary by dial-up or the like.

The term “entire user” includes not only all users but also a case where users are grouped and all users belonging to the group are indicated.

"Terminal device" means a personal computer, a mobile phone,
This is a concept including a mobile terminal and the like.

[0039]

BEST MODE FOR CARRYING OUT THE INVENTION FIG. 1 shows the configuration of a service system including a content recommendation system according to an embodiment of the present invention. The user terminals 12 and 14 are personal computers connectable to the Internet 4. Also, the user terminals 16, 18
Is a mobile phone connectable to the Internet 4. A management center 20 is connected to the Internet 4. It is preferable that the management center 20 be operated by a highly public organization such as a telephone communication company or a certified company thereof.

The management center 20 is a user authentication device 2
2, a usage recording device 24, a content recommendation device 26, an update time determination device 28, a recommended content notification device 30, and a service providing device 32. The service providing device 32 in the management center 20 is a so-called rental server, and is rented for a service provider who wants to open a web. In this embodiment, the service providing apparatus 32 has a weather forecast web 32a, a stock price web 32b, a sports news web 32c, a management web 32x, and the like by different service providers. Further, a service providing device 34 is provided outside the management center 20 and is connected to the Internet 4. A game web 36 is built in the service providing device 34.

FIG. 2 shows a hardware configuration of the user terminals 12 and 14. The user terminal 12 (14) has a memory 4
0, display 42, communication circuit 44, keyboard / mouse 46, CPU 48, hard disk (recording device) 5
0, a CD-ROM drive 52 is provided. The hard disk 50 stores an operating system (such as Microsoft Windows 98) and a browser program for browsing the web. This browser program is installed from the CD-ROM 54 via the CD-ROM drive 52. The communication circuit 44 is a circuit for connecting to the Internet 4.

User authentication device 22, usage recording device 24,
Content recommendation device 26, update time determination device 28, recommended content notification device 30, service providing devices 32 and 34
Are also the same as the hardware configuration shown in FIG. However, in the user authentication device 22, a user authentication program is recorded on a hard disk. Further, a usage recording program is recorded on the usage recording device 24, a content grouping program and a recommended content determination program are recorded on the content recommendation device 26, and an update timing determination program is recorded on the update timing determination device 28. The recommended content notification program is recorded in the content notification device 30.

In the service providing device 32, a web server program for the weather forecast web 32a, a web server program for the stock price web 32b, and a web server program for the sports news web 32c are recorded on the hard disk. . In the service providing apparatus 34, a web server program for the game web 36 is recorded on the hard disk.

FIG. 3 shows a user terminal 16 using a mobile phone.
18 shows a block diagram of FIG. As input / output devices, a liquid crystal display 62, a numeric keypad / switch 64, a microphone 6
6. Speaker (for call) 68, Speaker (for ringtone) 7
0 is provided. The audio encoding circuit 74 includes the microphone 6
6 is a circuit for encoding the voice from the speaker 6 for transmission, decoding the received voice signal, and outputting the decoded signal from the speaker 68. The micro browser 72 is a program recorded in the recording device, and is for browsing data from the service providing web. The wireless communication circuit 76 is a circuit for transmitting or receiving voice or data by wireless communication. The serial data communication circuit 78 is a circuit for performing communication with an external personal computer 84. The memory 80 stores the user's own telephone number, telephone directory, and the like. The control circuit 86 controls these circuits. The battery 82 supplies power to each unit.

In the system of FIG. 1, in order to receive a service from the service providing web, it is necessary to perform user registration in advance. FIG. 4 is a flowchart showing a program of the user terminal 16 and the user authentication device 22 for performing the user registration. Hereinafter, a case where user registration is performed from the user terminal 16 which is a mobile phone will be described.

First, the user terminal 16 accesses the user authentication device 22 via the Internet 4 (step S1). In response, the user authentication device 22 transmits a menu screen (step S11). User terminal 1
6, the menu screen is displayed on the liquid crystal display 62 by the micro browser (see FIG. 5A).

In this screen state, the user operates the numeric keypad / switch 64 of the user terminal 16 to set “2”.
Is input and the enter button is pressed (step S2). Thereby, the user authentication device 22 transmits a registration screen as shown in FIG. 5B to the user terminal 16. The user operates the numeric keypad / switch 64 of the user terminal 16 according to this screen, inputs a name, a zip code, a gender, a birthday, and the like, and presses a determination button. The input user information is transmitted to the user authentication device 22 (Step S3).

The user authentication device 22 acquires the subscriber ID of the user terminal 16 and uses it as the user ID. The subscriber ID is an identifier uniquely assigned by a communication company such as Daini Denden Co., Ltd. to the telephone number of the user. The user ID can be automatically obtained as data from a communication carrier such as Daini Denden Co., Ltd. by accessing the user authentication device 22 from the user terminal 16. The user authentication device 22 further generates a password, associates the password with the user information, and records the password on the hard disk as a user registration database. FIG. 6 shows a user registration database.
The information, ID, password, and the like of the registered user are recorded.

In this embodiment, the name, address, gender, occupation, credit card number and the like are recorded as registration items. In the present invention, these pieces of information are not always necessary, but are useful when a user group is generated based on these pieces of information, as described later.

After recording in the user registration database, the user authentication device 22 transmits a registration completion screen indicating the ID and the password to the user terminal device 16 (step S13). A registration completion screen displayed on the user terminal device 16 is shown in FIG. 5C.

The user registration is performed as described above. By performing the user registration in this manner, it becomes possible to receive the service provision of each content web by inputting the password corresponding to the user ID.

In the above description, the user terminal 16 which is a mobile phone has been described. However, registration is performed from the user terminal 12 which is a personal computer (connected to the Internet by a dial-up connection using a normal telephone or by a constant connection). In this case, the user authentication terminal device 22 cannot acquire the subscriber ID of the user terminal. Therefore, in the screen of FIG.
It is necessary to prepare a screen that allows the user to input an ID, and to perform processing such as allowing the user to input an ID.

Next, a process when a registered user uses a service (views content) will be described.

When the user authentication device 22 is accessed, a menu screen shown in FIG. 5A is displayed (step S2).
1). When "1" is selected on this screen, the user authentication device 22 transmits an authentication screen as shown in FIG. 5D (step S31). If the user terminal is a mobile phone, its ID (subscriber ID) can be obtained. Therefore, in this case, the user authentication device 22
D (see FIG. 5D), the user terminal 16
Send the screen to.

The user inputs the password from the user terminal 16 and presses the enter button (step S22). The user authentication device 22 determines whether or not the transmitted password matches the data registered in the user registration database (FIG. 6) (Step S32). If they do not match, it sends back a message that the passwords do not match. If the passwords match, a content menu as shown in FIG. 5E is transmitted (step S33).

From this content menu, the user can
The desired content is selected by number and the enter button is pressed (step S23). In response to this, the user authentication device 22 determines the ID of the content corresponding to the selected number.
Is transmitted to the usage recording device 24 (step S34).

In response to this, the use recording device 24 updates a use record database that records the number of times each content has been used for each user (step S41).

FIG. 8 shows the usage record database. The number of times each content is used is recorded for each user.
Note that the content name “A” corresponds to the weather forecast of FIG. 5E, “B” corresponds to the stock price of interest, and “C” corresponds to sports news. For example, if the user ID is 0125
01, the content A once, the content B eight times, and the content C during the target period (January 1 to 7).
Is displayed 0 times, the content D is viewed 5 times, the content E is viewed 0 times, and the content F is viewed twice. For example, if the weather forecast was selected by the user this time,
The number of contents A is updated to two.

Next, the web server program of the user authentication device 22 changes the browsing destination of the micro browser of the user terminal device 16 to the weather forecast web 32a by the redirection process (steps S35 and S24). Thereby, the user terminal device 16 is connected to the service providing device 32.
Can be received from the weather forecast web 32a (step S51).

As described above, every time a user browses the web, recording is performed by the user recording device 24,
A user record database is formed.

Next, at a preset update date and time,
The content recommendation device 26 performs a content grouping process. FIG. 9 is a flowchart of the processing program.
Shown in First, the content recommendation device 26 causes the user recording device 24 to transmit the user recording database at the update date and time (step S61).

Based on the usage record database, the content whose usage frequency is larger than the predetermined frequency is selected and grouped for each user (step S62).
For example, in the usage record database of FIG. 8, contents that have been used five or more times are grouped as shown in FIG. That is, the user “01250
For “1”, the group of BDs is
For "02", a BEF group is generated. If there is only one content exceeding the predetermined number of times for the user, a content group is generated using the one content.

The content recommendation device 26 records the content group generated in this way as a user-specific content group in association with the user ID and the usage points. Here, the usage point is the sum of the number of times of using each content of the group.

In FIG. 8 and FIG. 10, only three users are recorded. However, actually, many users are recorded.

Next, a content group for the entire user is generated based on the user-specific content group information (step S62). In this embodiment,
The usage points are totaled for all of the user-specific content groups (however, only those containing two or more contents). That is, when the same content group is generated by different users, the usage points of the different users are summed up to be the usage points of the content group.

For example, in FIG. 10, the content group BD includes the users “012501” and “01250”.
Since the number of use points is “3”, 13 and 10 are summed up. Furthermore, if the BD also appears for other users, the usage points of that user are also summed up.

As described above, the total usage points for all the content groups appearing in the user-specific content groups are calculated. Next, a content group whose total usage point exceeds a predetermined point is selected from the content groups, and this is set as a comprehensive content group (step S63).

FIG. 11 shows an example of the comprehensive content group information generated in this way.

The content group information for each user and the comprehensive content group information generated above are recorded in the content recommendation device 26 (steps S62 and S62).
63).

Next, a process for recommending content to each user by using the above-described user-specific content group information and comprehensive content group information will be described. FIG. 12 shows a flowchart of this processing.

The user terminal device 16 sends the user authentication device 2
The procedure from step S2 to step S2 to complete the user authentication is the same as steps S21, S31, S22 and S32 in FIG. When the user authentication is completed, the user authentication device 22 transmits the user ID to the content recommendation device 26 (Step S37).

The content recommendation device 26 receiving this,
The recommended content for the user is determined using the user-specific content group information and the comprehensive content group information (step S81).

FIG. 13 shows a process of determining a recommended content. First, a user ID is received from the user authentication device 22 (step S811). Next, the user ID is included in the recorded user-specific content group information (FIG. 10).
It is determined whether or not a content group related to the content is recorded (step S812). In the case of a user who has hardly used or a user who has just registered, the process proceeds to step S815 because no content group has been recorded. In step S815, the fact that there is no recommended content is returned to the user authentication device 22. In this case, some content may be recommended, for example, the content most used yesterday is recommended.

In response to this, the user authentication device 22 transmits a normal content menu that does not display recommended content to the user terminal device 16 (step S3 in FIG. 7).
3). FIG. 7 is a flowchart showing the processing in such a case, but the above-described determination part is omitted.

On the other hand, if the content group for the user ID is recorded in the user-specific content group information (FIG. 10), the flow advances to step S813. In step S813, it is determined whether there is a comprehensive content group that partially includes the content included in the user-specific content group of the user ID. If not, the process proceeds to step S815 as described above.

For example, it is assumed that the content group information for each user and the comprehensive content group information as shown in FIGS. 10 and 11 are recorded, and the user ID “012501” is accessed. In this case, the content B included in the user-specific content group BD,
Comprehensive content group 5 (see categorization number in FIG. 11) that partially includes D can be found.

When such a comprehensive content group is found, the content included in the comprehensive content group but not included in the user-specific content group is set as the recommended content (step S814).
In the above example, the content C is determined as the recommended content.

The reason for determining the recommended content in this way is as follows. Comprehensive Content Group 5
Contains contents B, C, and D. That is,
It is shown that many users use the contents B, C, and D collectively. On the other hand, the target users frequently use the contents B and D. From the comprehensive content group information, it can be determined that C is worth watching for those who use B and D. Therefore, the recommended content is determined based on the above determination.
Note that a plurality of contents may be recommended contents.

Content ID of determined recommended content
Is transmitted to the recommended content notification device 30 (FIG. 1).
2. Step S81). The recommended content notification device 30 receiving this transmits a content menu with recommended content to the user terminal device 16 based on the content ID (step S91). In the user terminal device 16, a content menu with recommended content is displayed as shown in FIG. 5F (step S9). The subsequent processing is the same as the processing after step S23 in FIG. As described above, content suitable for each user can be recommended.

According to this embodiment, the content is automatically grouped according to the usage status, and the content is recommended for each user. Therefore, there is no need for the administrator to maintain grouping, and individual content recommendations can be made for each user.

It is preferable that the grouping process in the content recommendation device 26 be recalculated every predetermined period. For example, the grouping may be recalculated every fixed period, such as every seven days. In this embodiment, an update time determination device 28 is provided,
The recalculation time is dynamically changed according to the situation.

The update time determination device 28 compares the previous comprehensive content group with the current comprehensive group, and determines the organization until the next recalculation according to the magnitude of the change. FIG. 14 shows this state. On January 1-7, when the recording started, 1
Four comprehensive content groups 2, 3, and 4 have been generated.

At the start of the next period (8 to 14 days), 7 days, which is the default update period, is set.
In the period from the 8th to the 14th, three comprehensive content groups 1, 2, and 4 are generated. That is, the content is reduced by one.

The update time determination device 28 calculates the increase or decrease of the content as a variation coefficient. Here, the variation coefficient is “1”. Further, the update time of the next period (15 days to 21 days) is determined based on the coefficient of variation. The determined update time is transmitted to the content recommendation device 26.
The content recommendation device 26 executes updating (recalculation) of the next grouping according to this.

That is, when the variation coefficient is large, the update time is short, and when the variation coefficient is small, the update time is long.

Thus, according to the magnitude of the fluctuation,
Update timing can be dynamically changed. This makes it possible to maintain the latest grouping that appropriately responds to changes in the situation while minimizing the burden of updating.

2. Usage Points in Usage Recording Device 24 In the above embodiment, each time a user uses the content once, one point is given to the content. However, weighted points may be given according to the content browsing status. For example, 1 point can be set for content that can be viewed free of charge, 3 points can be set for content that is viewed for a fee, and 5 points can be set for content that involves purchase of a product or the like. By doing so, it is possible to give high points to important browsing activities according to the browsing status even in the same one browsing.

Further, points may be given in accordance with one browsing time. For example, 0 for less than 1 minute
1 point for 1-5 minutes, 2 for 5-10 minutes
Points are given, 3 points for 10 to 30 minutes, and 5 points for 30 minutes or more.

Further, high points may be given to the repeat browsing. For example, if the user has viewed the content A, and the user has viewed the content A last time within three days, the user can obtain 1.2 times the point when the user last viewed the content A. Points. Therefore, a high point is given to content that is frequently viewed repeatedly in a short period of time.

Further, the points given for browsing the content may decrease as the number of days elapses. For example, as shown in FIG. 15, a parameter t is set according to the number of days that have elapsed since the browsing, and points multiplied by t are adopted as points. In this way, a higher point is given to a more recent browsing, and a recent tendency can be appropriately reflected.

Note that the methods for calculating each point described above may be used in combination with each other.

3. Content recommendation device 26 In the above embodiment, a content group for each user is generated based on the content whose content points within a predetermined period have exceeded a predetermined value, and a comprehensive content group is generated based on the content group. . However, a user-specific content group and a comprehensive content group may be generated as follows.

(1) Method of Generating User-Specific Content Group for Each Use First, based on the usage record from the usage recording device 24, the user can use the content group within a predetermined period (for example, one day). The (viewed) contents are grouped. Thereby, as shown in FIG. 16A, a user-specific content group is obtained for each day. That is, in this figure, ABC, D
Four user-specific content groups E, AB, and EFG are obtained.

Next, points are given to each user-specific content group in consideration of the number of contents included in the group, repeat browsing as a group, and the like (see FIG. 16B). In this embodiment, it is not necessary to give points in the usage recording device 24.

The above processing is performed for all users to generate user-specific content groups for all users.

Next, user-specific content groups of all users are put together. For example, if the same content group is generated for two or more users, this is regarded as one content group, and the points of each user for the content group are totaled. Let the total points be the points of the content group. In this way, a content group and its points as shown in FIG. 16C are obtained.

Next, from this content group, those having a predetermined point (for example, 10,000 points) or more are selected, and this is set as a comprehensive content group. FIG.
In the case of C, DE and AB are selected as the comprehensive content group.

As described above, a user-specific content group and a comprehensive content group can be generated. In the method shown here, a set of contents used during a predetermined period (one day in the above example) of each user is set as the user-specific content, so that more appropriate user-specific content can be formed.

The above method may be performed with one access (from connection to disconnection) as a unit of a predetermined period.

In the above description, data as shown in FIG. 16A is obtained based on the usage record of the usage recording device 24. However, the data as shown in FIG. 16A may be recorded in the usage recording device 24 as a usage record.

(2) Method of Generating Content Group by User Based on Average Points First, the average of the points of each content within a predetermined period is calculated for one user. Next, a content group for each user is generated based on the content in which points having a predetermined magnification α (for example, 1.5 times) of the average points are recorded. For example, for a usage record as shown in FIG. 17, 7.5 (average point) × 1.5
Content whose content exceeds 11.25 points is selected as a user-specific content group for this user. A comprehensive content group is generated based on the thus generated user-specific content groups for each user.

A plurality of predetermined magnifications α may be provided, and a user-specific content group and a comprehensive content group may be generated according to each predetermined magnification. In this case, the higher the magnification α, the higher the association between the contents. Therefore, as described later, it is possible to preferentially recommend contents belonging to a comprehensive content group having a high magnification α.

(3) Method of Generating Content Group Using Correlation Function For each user, a content group for each user is generated based on the usage record. Any of the above-described methods can be used as the generation method.

Next, for all users, a correlation coefficient is calculated for certain two contents from the viewpoint that "the point of one content is high, the point of the other content is high". If the correlation coefficient exceeds a predetermined value, the two contents are selected as a comprehensive content group.

Similarly, the same correlation coefficient as described above is calculated for three or more contents to generate a comprehensive content group.

When such a correlation coefficient is used, the degree of association between contents may be recorded as a network of correlation coefficients as shown in FIG.
Then, for a user who has viewed only one of the two contents having a high correlation coefficient, the other content can be recommended.

(4) Addition of Temporal Element to Content Group Even if the same content is used, it may be handled as different content for each time zone (for example, morning and afternoon), and a content group may be generated. . Similarly,
The content may be treated as different depending on the day of the week. By doing so, for example, it is possible to say, "For those who watch the content A on Monday, the content B is recommended on Friday".

(5) Grouping of Users In each of the above embodiments, a comprehensive content group is generated for all users, and recommended contents are determined. However, the users may be grouped by some attribute, and a comprehensive content group may be generated for all users belonging to each group. In this case, a comprehensive content group is generated for each user group. Also,
When determining recommended content for a user, a content group for each user of the user and a comprehensive content group of the user group to which the user belongs are used. In this way, according to the attribute of the user,
More appropriate content recommendations can be made.

The attributes for grouping users include i) the residence area of the user, ii) the district where the user used the terminal device, iii) the gender of the user, iv) the age group of the user, v)
The occupation / occupation of the user, and vi) the type of terminal device used by the user (eg, personal computer, PDA, mobile phone, PHS, etc.) may be used.

(6) Recommendation Considering Content Transition In a comprehensive content group, grouping is performed in consideration of the browsing order of the content included in the group. For example, as shown in FIG. 19, a comprehensive content group including the tendency of the browsing order is generated. In this case, the content A is recommended to the user who has viewed the content F.

4. Recommended Content Notification Device 30 In the above embodiment, when there is a recommended content, the recommended content notification device 30 generates a portal site and notifies the portal site. However, a recommendation by a banner advertisement, a recommendation by an e-mail, a recommendation by a message push, etc. may be used.

When there are a plurality of recommended contents, the contents having higher priority may be displayed with priority. For example, displaying only the recommendation of the high-priority content, displaying the higher the recommendation of the higher-priority content, displaying the higher-priority content recommendation from the lower-priority content recommendation each time the switch button is pressed, and the like. Can be. This priority can be determined by the above-described magnification α, correlation coefficient, and the like.

[0113] 5. Others In the above embodiment, the recommended content is determined based on the user-specific content group and the comprehensive content group. However, instead of using the user-specific content group, the recommended content may be determined each time the content is used, using the information of the used content and the comprehensive content group.

Further, in the above embodiment, the usage record is taken for each user, and the recommendation is made. However, on the premise that the same terminal device is used by the same user, usage records may be taken for each terminal device and recommendations may be made.

In the above embodiment, the user authentication device 22, the use recording device 24, the content recommendation device 26, the update time determination device 28, and the recommended content notification device 30
Configured as separate devices. However, some or all of these devices may be collectively operated as one device.

[Brief description of the drawings]

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

FIG. 2 is a diagram showing a hardware configuration of a user terminal 12, a user authentication device 22, and the like.

FIG. 3 is a diagram showing a hardware configuration of a user terminal 16;

FIG. 4 is a diagram showing a flowchart of user registration.

FIG. 5 is a diagram showing a screen display on the user terminal device.

FIG. 6 is a diagram showing a user registration database.

FIG. 7 is a diagram showing a flowchart of browsing by a user.

FIG. 8 is a diagram showing a usage record database.

FIG. 9 is a diagram illustrating a flowchart of a grouping process.

FIG. 10 is a diagram showing user-specific content group information.

FIG. 11 is a diagram showing comprehensive content group information.

FIG. 12 is a diagram showing a flowchart of browsing by a user.

FIG. 13 is a diagram showing a flowchart of content recommendation.

FIG. 14 is a diagram showing a fluctuation state of a group.

FIG. 15 is a diagram showing parameters indicating the latest degree of browsing (use).

FIG. 16 is a diagram illustrating a method of generating a content group according to another embodiment.

FIG. 17 is a diagram illustrating a method of generating a content group according to another embodiment.

FIG. 18 is a diagram showing an association between contents represented by a network.

FIG. 19 is a diagram showing comprehensive content group information in consideration of content transition.

[Explanation of symbols]

 4: Internet 12-18: User terminal 20: Management center 22: User authentication device 24: User recording device 26: Content recommendation device 28: Update time determination Apparatus 30 ・ ・ ・ Recommended content notification apparatus 32 ・ ・ ・ Service providing apparatus

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Shinichiro Araki 5-22 Higashino Kitainoue-cho, Yamashina-ku, Kyoto City, Kyoto Prefecture Inside Kyocera Communication System Corporation (72) Inventor Hiroshi Yoshida 2-14 Tamagawadai, Setagaya-ku, Tokyo -9 Kyocera Communication System Co., Ltd. Tokyo Branch Office F-term (reference) 2C001 AA00 AA17 BB00 BB01 BB05 BD00 BD04 BD07 CB01 CB06 CB08 CC02 5B075 KK03 KK07 KK13 KK33 KK37 KK43 KK54 KK63 MM11 ND03 NK10 NK23 PP13 PQ02 PQ04 PQ32 PQ75 PR08 UU24 UU40 9A001 JJ25 JJ76 KK62

Claims (19)

[Claims] A terminal device for browsing content, a plurality of content providing means for providing content to the terminal device via a network in response to a request from the terminal device, and a user of the terminal device browsing. Usage status recording means for recording content for each user, and usage status for each user recorded in the usage status recording means,
A related acquisition unit for acquiring a relationship between contents based on the usage status totalized for all users, and a usage status of each user recorded in the usage status recording unit or information derived from the usage status. A content recommendation system comprising: a content recommendation unit that determines recommended content for each user based on the inter-content related information generated by the acquisition unit, and recommends the recommended content to the user of the terminal device. 2. A content recommendation device connected to a terminal device via a network and for recommending content to a user of the terminal device, wherein usage status of the content for each user is totaled for all users. Based on the relationship acquisition means for acquiring the relationship between the contents based on the use situation, the use situation of each user or information derived from the use situation, and the content-related information generated by the relationship acquisition means, A content recommendation device comprising: content recommendation means for determining recommended content for each user. 3. The system according to claim 1, wherein the content-related information is grouping information obtained by grouping each content. 4. The system or apparatus according to claim 3, wherein the related acquisition unit groups related contents for each user based on the usage status of each user recorded in the usage status recording unit. And generates comprehensive content group information that groups related content for the entire comprehensive user as grouping information based on the user-specific content group information. The content recommending means determines recommended content for each user based on user-specific content group information and comprehensive content group information. 5. The system according to claim 1, wherein the inter-contents related information is information indicating a degree of relevance between the respective contents. 6. A content recommendation method in a system in which a plurality of contents of one or more content providing devices can be used in a terminal device via a network, wherein a usage tendency of each content is acquired for each user. , Without discriminating which users use the content, obtain comprehensive usage trends of each content, and compare the comprehensive content usage trends with each user's individual content usage trends. Deciding which content should be recommended for each user. 7. A terminal device for browsing content, a plurality of content providing means for providing content to the terminal device via a network in response to a request from the terminal device, and a user of the terminal device browsing. Usage status recording means for recording content for each user, and user-specific content in which related content is grouped for each user based on the usage status for each user recorded in the usage status recording means Related acquisition means for generating group information and generating comprehensive content group information in which related content is grouped for the entire user based on the content group information for each user or the usage status for each user; User-specific content group information and comprehensive content group information generated by related information acquisition means And a content recommendation means for determining recommended content for each user based on the information and recommending the recommended content to the user of the terminal device. 8. A content recommendation device connected to a terminal device via a network and for recommending content to a user of the terminal device, wherein the content recommendation device is based on a usage status for each user recorded in a usage status recording means. , Generating user-specific content group information in which related content is grouped for each user, and relating the entire user based on the user-specific content group information or the usage status of each user. Content group generating means for generating comprehensive content group information in which the content is grouped, and user-based content group information and comprehensive content group information generated by the content group generating means, for each user, Determine recommended content and recommend it to terminal device users Content recommendation apparatus comprising: means recommend, the. 9. The system according to claim 7, wherein the use status recording means records the number of times each content is used for each user, and wherein the content group generation means includes A feature of obtaining user-specific content group information based on the number of times content is used. 10. The system or apparatus according to claim 9, wherein said usage status recording means records usage points in consideration of weighting based on usage types, and said content group generation means includes a usage point of each content. A method of obtaining user-specific content group information based on 11. The system or apparatus according to claim 7, wherein said use status recording means records the use time or day of each content for each user together with said content group. The generation means obtains comprehensive content group information for each use time zone or day of the week, and the content recommendation means obtains comprehensive content group information for the time zone to which the time belongs or the day of the week according to the current time or the day of the week. The recommended content is determined for each user using. 12. A system or apparatus according to claim 7, wherein the same content is treated as different content for each time zone or day of the week. 13. The system or apparatus according to claim 7, wherein said content group generating means comprises: a user-specific content group and a comprehensive content group;
Grouping is performed in consideration of the browsing order of each content, and the content recommending unit also considers the browsing order,
Determining recommended content for each user. 14. The system or apparatus according to claim 7, wherein said content group generating means obtains comprehensive content group information for each attribute group based on a user attribute, and said content recommending means. Is based on the user-specific content group information and the comprehensive content group information of the attribute group to which the user belongs,
Determining recommended content. 15. The system or apparatus according to claim 7, wherein the content recommending means finds a user-specific content group constituted by a part of the content included in the comprehensive content group, and uses the content group. Content that is not included in the personalized content group but is included in the comprehensive content group
Characteristically determined as recommended content. 16. The system or apparatus according to claim 7, wherein said content group generating means recalculates user-specific content group information and / or comprehensive content group information at predetermined time intervals. What is characterized by. 17. The system or apparatus according to claim 16, wherein said content group generation means determines the timing of the next recalculation based on the previous comprehensive content group information and the degree of change of the current comprehensive content group information. Characterized by a decision. 18. A content recommendation method in a system in which a plurality of contents of one or more content providing devices can be used in a terminal device via a communication line, wherein the method is frequently used for each user or terminal device. Content is grouped for each user to obtain user-specific content group information, and for all users, commonly used content is grouped to obtain comprehensive content group information. , Comparing with comprehensive content group information,
Determining which content should be recommended, a content recommendation method. 19. A method of grouping a plurality of contents according to the use status, wherein a plurality of contents used by a same user within a predetermined period are set as a temporary group, and a plurality of contents are grouped for a plurality of predetermined periods. The temporary group is generated, and based on the number of uses and the type of use of each content included in each of the temporary groups for all users,
The usage points are calculated for each temporary group, and only the temporary groups that have obtained usage points exceeding a predetermined value are
A grouping method, characterized in that: