JP2001209660A - Contents retrieval/recommendation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system for efficiently retrieving/recommending new products suited to the preferences of individuals in the contents market of music or the like where many new products are put on the market daily for demanders having wide preferences, developing contents for satisfying latent needs by actualizing the latent needs of users and supporting the presentation of the contents with potential to latent customers. SOLUTION: A control center 20 is provided with a contents data base 27 and a contents parameter data base 26 for storing the sensuous features of the contents as a parameter set 16 which is the set of quantized attributes. A retrieval terminal 33 is provided with a means for previewing sample contents every time a retrieval condition is set to the parameter set. After the utilization of the contents, the set value of the parameter set is evaluated in the retrieval terminal and the set value is corrected in the control center.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a variety of contents, such as new music, which are unknown to users, in areas such as online interactive networks and digital multi-channel broadcasting. The present invention relates to a system for efficiently and effectively searching and recommending high-content.

[0002]

BACKGROUND OF THE INVENTION In addition to the conventional "package media" and "broadcast media", so-called "on-line interactive media" such as the Internet has appeared by digital technology. The width has expanded dramatically.

Some technologies for obtaining desired information from such a so-called "flood of information" have begun to be developed, and these existing technologies are roughly characterized by the following three types.

[0004] 1. Assuming complete information state: This is based on the premise that the attributes to be searched are fixed, the searcher has knowledge of them, and the searcher's preferences have already been determined a priori. Originally, the search target and the searcher are systematically linked.

[0005] In many of these systems, the characteristics of the search target are characterized by keywords or the like. If there is universal knowledge (keywords or the like) about the attributes of the search target, an effective search such as the author or title is performed. It is possible to search for books and songs by name.

[0006] However, even when the searcher has known the search target, has already seen, heard, and read the search object, and has given high evaluation to the search object, the searcher is familiar with such keywords. Without it, you can't do an effective search. As an example of such a case, for example, there is a case where the user likes the song heard on the radio but does not know the title.

[0007] Further, when the search target is unknown to the searcher, and when a sensuous object such as music or an image is handled, in a fixed keyword search, a narrowed search is performed by combining keywords. However, it is not easy to reach content with a high degree of satisfaction.

[0008] 2. Guidance for selecting searchers: This is for searchers (beginners) with little knowledge of the attributes (keywords, etc.) to be selected.
It is intended to add a “guidance function” in attribute selection. The simplest example is a method in which a searcher specifies a specific favorite category in advance, and preferentially introduces related items.

[0009] However, such guidance is based on a fairly wide set of predefined sets (in the case of music, rock, jazz, classical, etc.), and is therefore a recommendation for individuals with diverse tastes. Is likely to be uniform. In addition, for a searcher who listens to such genres in a cross-cutting manner, a considerably large number of genres are specified, and as a result, desired contents are not narrowed down. In fact, in the case of music, the number of lovers who listen to music across genres has increased dramatically compared to the past.

[0010] 3. An attempt to predict and recommend content with a high degree of satisfaction from individual searchers' preference evaluation information for individual content: These systems have recently attracted attention, and have the image In order to dispel oneness, by collecting and analyzing searchers' preferences widely, searchers with similar preferences can be grouped, and the content with a relatively high evaluation in the group can be replaced with other content. This is a system that tries to recommend to group members.

[0011]

However, these conventional systems have the disadvantages that (A) content with a wide variety is already available in the market in all or a sufficient amount to satisfy the tastes of searchers. (B) It is based on the premise that it is easy to collect as much searcher preference data as possible for such a huge amount of content to lead to a statistically significant solution.

[0012] However, in the real world, these two assumptions are far from being true.
(A ') New contents continue to be supplied to the market every day, and it is difficult to grasp the whole at one time. (B') All contents including known ones are statistically significant. It is not easy to collect the searcher's preference data from the beginning enough to lead to a proper solution.

Further, even if there is an environment for collecting statistical information that satisfies the above condition, only the evaluation information for the existing contents can be analyzed, so that the potential needs of the user cannot be known.

Therefore, aside from the theory, in reality, in a variety of markets such as a music content market in which a large number of new works are put on the market every day to consumers having a wide range of tastes, new works that match individual tastes are realized. By efficiently searching for and recommending content, and revealing the potential needs of users, leading to the development of content that satisfies those potential needs, and presenting potential content to potential customers A system to help with this has not yet been developed.

[0015]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention searches for an efficient content search function from the beginning even when it is difficult to immediately obtain preference correlation data based on evaluation data of a considerable number of searchers. The first purpose is to provide to the person.

A second object of the present invention is to make derivation easy,
In addition, it is possible to recommend content with high significance.

A third object of the present invention is to provide a system capable of eliciting latent demand for content and deriving marketing information relating to content production / supply later.

[0018]

In order to achieve the above object, according to the first aspect of the present invention, contents stored in a control center are used by a search terminal connected to the control center via a communication line. The system includes a content center that stores content and a content parameter database that stores sensory characteristics of the content as a parameter set that is a set of one or more quantified attributes. The search terminal performs a search operation using a parameter database, receives a recommendation of a content group from the content database in accordance with a search condition, and makes a desired content available from the content group.

According to a second aspect of the present invention, in the content search / recommendation system according to the first aspect, the search terminal displays a display and a parameter setting screen on which a search condition can be set for the parameter set. Means for displaying an operation lever capable of designating a search condition having an allowable range for each attribute of the parameter set by sliding the parameter setting screen, and an operating lever according to the allowable range. Means for visually displaying a designated width for designating a search condition on a parameter setting screen.

According to a third aspect of the present invention, there is provided the content search / recommendation system according to the first or second aspect, wherein the control center further determines the sample content corresponding to the set value of the parameter set. Means, wherein the search terminal further comprises means for previewing sample content corresponding to the search condition every time a search condition is set in the parameter set.

According to a fourth aspect of the present invention, there is provided the content search / recommendation system according to the first or second aspect, wherein the content database is a database for storing music content, and the control center further comprises a parameter set. The search terminal includes sample content determination means for determining sample music corresponding to the set value,
Further, each time a search condition is set in the parameter set, a means for reproducing the sample music corresponding to the search condition is provided.

According to a fifth aspect of the present invention, there is provided the content search / recommendation system according to any one of the first to third aspects, wherein the search terminal further modifies parameters for each content after the user uses the content. The control center further includes means for displaying an evaluation screen on which a request can be input, wherein the control center further collects parameter correction request information input on the evaluation screen via the communication line, and sets the parameter set values. And a parameter setting means for changing the parameter.

According to a sixth aspect of the present invention, there is provided the content search / recommendation system according to any one of the first to third aspects and the fifth aspect, wherein the control center further executes a search process of the search terminal. A user behavior data management unit that collects information for each user identified by the predetermined determination unit, a user behavior database that stores information collected by the user behavior data management unit, and a user behavior data stored in the user behavior database. A recommended content determining unit that determines recommended content for each user from the setting information of the content parameter.

According to a seventh aspect of the present invention, there is provided the content search / recommendation system according to the sixth aspect, wherein the recommended content determining means further comprises a user who has previously used the search result specified by the user. And a means for omitting the content.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A preferred embodiment of a content search / recommendation system will be described below with reference to the drawings, taking a music content preview system as an example.

<Music Content Search / Recommendation System> FIGS. 1 and 2 show the overall configuration of an embodiment in which the present invention is applied to a music content search / recommendation system. The system of this embodiment includes a record company 11, a sound source provider 12,
Content data provider 10 including record association 13
And a control center 2 having a content parameter database 26, a content database 27, etc.
0, a service using unit 3 for receiving services provided by the content data providing unit 10 and the control center 20
The service use unit 30 includes a store 31 and an individual user 32.

The content data providing unit 10 and the control center 20 are connected by the communication line 1.
The control center 20 and a search terminal 33 provided in each store 31 are connected by a communication line 2, and the control center 20 and a search terminal (PC) 34 owned by each individual user 32 are connected to the communication line 3. Connected by

The search terminal 33 of each store 31 and the search terminal (personal computer) 34 of each individual user 32 basically have the same function, but the search terminal 34 of the individual user 32 Realized on personal computers at home,
It is assumed that the service can be easily provided from a browser or the like.
Is preferably set to a minimum necessary configuration. In the following description, the search terminal 33 of the store 31 is mainly
Although the search terminal is described as a search terminal of 0, the same configuration can be applied to the search terminal 34 of the individual user 32.

The form of each of the communication lines 1, 2, 3 is not particularly limited, and may be any configuration such as the Internet, a telephone line (ISDN line), a dedicated line, a cable television network, and the like.

The trial sound source data 14 and the accompanying information 15 provided from the content data providing unit 10 are transmitted to the control center 20 through a computer network (communication line 1) or a package medium such as a CD-ROM. Note that the audio-visual sound source data 14 is transmitted to the audio data compression unit 23, and the supplementary information 15 including character information such as a music title and an artist name and still image information such as a jacket photograph is transmitted to the supplementary information input unit 22. .

The audio data compression section 23 performs data compression processing of sound source data in order to perform network distribution more smoothly, and the supplementary information input section 22 formats the supplementary information 15 such as music titles and performers into a common format. Such information is stored in the content database 27 in the control center 20.

However, the audio data compression section 23 and the content database 27 need not necessarily be located in the control center 20. The content data providing unit 10 controls the control center 2 due to copyright considerations.
In the case of reluctance to store the sound source within 0, the store 31 described later
Alternatively, content data distribution to the individual user 32
The content data providing unit 10 may perform the processing directly.

Conversely, with the permission of the owner of the content copyright, each time the service user searches for the content, the content is supplied from the content data supply unit 10 or the content database 27 of the control center 20 to each store 31. Instead of distributing the content data to the (search terminal 33), the content data may be stored in the search terminal 33 in the store 31. By constructing a system capable of local processing at the store 31 in this way, it is possible to suppress an increase in running costs mainly caused by communication costs between the search terminal 33 and the control center 20.

As described above, the content data (the preview sound source data 14 and the supplementary information 15) are directly distributed by the content data providing unit 10, and the control center 2
Both the configuration in which the content database 27 in 0 stores and distributes and the configuration in which each store 31 stores are possible. It should be noted that although a configuration in which the content database 27 is stored in each individual user 32 is technically possible, the personal database 32 uses a personal computer that can be connected to the Internet by a modem, for example. Because an unspecified number of users are assumed,
It is preferable to limit data to be stored.

On the other hand, the content parameter database 26 in the control center 20 is a core part of the content search / recommendation system. The parameter set 16 for each content set in advance by the content parameter setting unit 21 is stored in the content parameter database 26.

The parameter set 16 expresses the sensory characteristics of the content as a set of one or a plurality of quantified (numerical) attributes. Each content has the characteristics of the content. It can be searched by expressing it by the parameter set 16.

The recommended content determining unit 28 includes a search terminal 33 of the store 31 connected by the communication line 2 such as a dedicated line or a search terminal (personal computer) 34 of the individual user 32 connected by the communication line 3 such as the Internet. The recommended content is extracted by the procedure described later using the content parameter database 26 in accordance with the instruction from
The list of the recommended contents is fed back to the third class or the like. These recommended content lists are stored in each search terminal 33.
And so on, the searcher can use the control center 20
A desired content can be previewed by means such as download or stream distribution directly from the content database 27 or from the content data providing unit 10 as described above.

After previewing the content, each service user is asked for an opinion on the parameter set 16 of the content set in advance, and if there is any objection to the parameter set 16 set by the control center 20, The parameter set 16 considered to be valid is notified to the control center 20 side.

The information is collected by the parameter setting section 21 in the control center 20, and if necessary, the parameter set 16 of the corresponding content data is obtained.
To change. By doing so, it is possible to reduce the possibility that the parameter set 16 preset by the control center 20 becomes self-righteous.

A mirror database 42 of the parameter database 26 is installed in each store 31, and updated information such as the parameter set 16 of a new musical score is transmitted to the parameter setting unit 21.
May be transmitted to the store 31 side. The transmission timing may be real-time, or may be transmitted every fixed time or every time a certain amount of data is stored.

When the mirror database 42 of the parameter database 26 is installed in each store 31, each service user does not need to access the content parameter database 26 in the control center 20 every time the search is performed. Cost can be reduced.

In the case where the trial content is easily stored in the store while the communication cost is relatively high, it is most desirable that the content parameter database 26 and the content database 27 be stored in the control center 20 and the store 31 be stored. Side mirror database 42
Is to build a hybrid system that also accumulates data. Frequently accessed information is local (store 31
Side), and using information that is not usually used so much by using the center access can reduce the communication cost and improve the actual performance of the search.

The sample content determining section 29 is a means for determining a sample content from a plurality of contents that match the search conditions set in the search terminal 33. Then, before making a search request, the user can listen to the sample music corresponding to the search condition on the search terminal 33. By providing such a function, the user can make a search request after checking the validity of the search condition to some extent while listening to the sample music, thereby reducing unnecessary search access,
Work efficiency can be improved.

Lastly, the user behavior data management unit 24 manages the content parameter set 16 collected by each search terminal 33 and 34 by the individual service user at the time of the search, and the individual content after the content preview. The obtained evaluation information 17 is downloaded to the control center 20 through the communication lines 2 and 3 and stored in the user behavior database 25.

The parameter set 16 must be stored as time-series data for each individual searcher. In this case, however, it is not necessary to collect the searcher's own profile (name, gender, address, etc.). It is only necessary to confirm whether or not they are the same person by the card ID or the like.

Further, the evaluation information 17 is not collected as time-series data for each searcher, but is collected for each parameter set set by each searcher at the time of search or for each parameter set of the content corrected after listening. It may be just that. Therefore, the customer database 41 and the like dealing with the customer profile of each store 31 are installed in the individual store 31 and are not managed by the control center 20.

As described above, in the present embodiment, since unlimited transmission of personal information between the control center 20 and the store 31 (the search terminal 33) is not involved, the risk of information leakage relating to personal privacy infringement is minimized. Can be stopped.

In this way, the recommended content determination unit 28 can only recommend a content having a relatively high expected value of satisfaction from the evaluation information 17 evaluated for each parameter set even for a user who does not notify the searcher ID. Rather, for the searcher who has announced the searcher ID, the content that has already been previewed is omitted from the introduction content, and a group having the same parameter setting tendency is identified from the parameter setting tendency information of each searcher, Content with a high evaluation value in the group can be individually recommended for each searcher.

<Structure of Content Parameter Database in Music Content Search / Recommendation System>
The format of the parameter set 16 used in the content parameter database 26, which is the core of the content search and recommendation system, will be described. FIG. 3 is a diagram schematically showing a setting example of the parameter set 16 of music content. Each parameter expresses a general-purpose “characteristic” characterizing each music content. It is relatively expressed by the level of quantified numerical values (including negative values).

In this example, a music piece is defined by a combination of corresponding values in the range from “+10” to “−10” regarding six parameters (attributes), such as attitude, key, rhythm, performance, musical instrument, and recording. ing. On the same parameter database 26, each music piece has the same parameter set 16 according to each characteristic.
Are defined by a combination of different numerical values.

The parameter set 16 taken up in the example of FIG. 3 is an example, and a parameter database using another parameter set may be created. However, as the parameter set 16 constituting the content parameter database 26, By changing the value of each parameter,
It is desirable to set a parameter group that allows all contents to be expressed gradually and comprehensively.

For example, in the parameter "key", all songs can be positioned gradually between the opposite attributes of major = "bright" and minor = "dark". If there is a song that you can not feel the tonality,
This means that this parameter is “0”. Generally, by using a sensory, quantitative or continuous parameter such as “tone”, many contents can be classified and expressed.

On the other hand, if a discontinuous parameter such as "genre" is set and the extreme values are positioned as "rock" and "classic", the "genre" of "jazz" which does not belong to either of them. Must be defined as medium content. If such a definition is made, an attribute axis (= parameter axis) of another dimension must be set to complement them, so that the number of parameter sets inevitably increases. Therefore, it is easier to provide simpler and more effective guidance without relying only on such discontinuous parameter settings.

FIG. 4 shows an example of a data record for each song in the content parameter database 26.
Songs 1 to 5 are defined by combinations of corresponding values of six parameters, such as attitude, key, rhythm, performance, musical instrument, and recording. In this example, each parameter is defined as an integer value in a range from “−5” to “+5”, and the meanings thereof are as follows.

* "Attitude": "+" side = "avant-garde", "-" side = "conservative" * "Key": "+" side = "bright", "-" side = "dark" * "Rhythm": "+" side = fast tempo, "-" side = slow tempo * "Performance": "+" side = "dense" arrangement, "-" side = "rough" Arrangement * "Instrument": "+" side = used instrument is mainly "Electric", "-" side = used instrument is mainly "Acoustic" * "Recording": "+" side = "Hi-""Fi" recording, "-" side = "Lo-Fi" recording.

A general genre is described in a column 16a next to the six parameter columns of the parameter set 16. If there are too many tunes even after narrowing down the content parameters, the content can be narrowed down more efficiently by specifying parameters after filtering with general genres.

The last column 16b is an industry standard music ID of the music set by the record association or the like.
Used when called from another database or the like that adopts the D code.

<User Interface of Search Terminal in Music Content Search / Recommendation System>
Search terminal 3 in music content search and recommendation system
The user interface 3 (34) will be described. FIG. 5 is a configuration and an overview of the store search terminal 33.

The personal computer 51 installed in the back yard of the store 31 is connected to the control center 20 by the communication line 2, and furthermore, the monitor cable and the RS
It is connected to a display 52 with a touch panel at a store by a 232C cable or the like.

The display 52 with the store touch panel
Is provided with a magnetic card slot 53, and the service user can notify his / her own ID to the control center 20 side by inserting his / her magnetic point card or the like into the slot.

A speaker 5 for listening to the searched music is provided on the side of the display 52 with a touch panel.
4 are provided. This may be replaced by headphones or the like. The personal computer 51, the display with a touch panel 52, the magnetic card slot 53 and the speaker 54 constitute the storefront search terminal 33.

Next, the buttons arranged at the top of the screen are:
Genre setting button 5 for "pre-filtering"
5 (shown in an enlarged state in FIG. 6). "Pre-filtering" refers to narrowing down the search target to some extent in advance by using information such as existing genres before interactively narrowing down the value of the content parameter on the content parameter setting screen. . That is, this is a narrowing-down search by the parameter shown in the column 16a in the example of the parameter data record shown in FIG.

The buttons include “World” and “Roc”
genres such as "k" and "J-POP" are displayed.
When these are touched, the contents displayed in the recommended content list display section 56 thereunder are narrowed down to those belonging to the genre, and at the same time, the designated genre is stored in the pre-filtering information storage section of the search terminal 33. It is memorized.

Here, since there is a limit to the genre that can be displayed as the genre setting button 55, one of the buttons is "Other" (the button indicated by 55a in the figure).
When this button is touched, a pull-down menu is displayed, and it is preferable that other genres can be selected.

One of the genre setting buttons 55 is “recommended by our shop” (button indicated by 55b in the figure).
May be set, and the recommended content determined in the personal oriented recommendation process described later may be displayed by touching the button.

Next, a button displayed below the recommended content list display section 56 is a parameter setting screen display button 57. When this button 57 is touched, a parameter setting screen 60 is displayed. The section below it is the advertising section. The title to be advertised is displayed, and a description of the title can be displayed.

Next, FIG. 6 shows a parameter setting screen 60 displayed when the button 57 is touched. By moving up and down the levers 61 displayed in the center (six in this embodiment), the characteristics of the search music shown in FIG. 3 can be designated. These may be represented by an image of an operation device such as a dial instead of the lever 61.

Also, in this terminal, every time the lever 61 on the screen is operated, a "standard music sample" represented by the currently designated parameter is played. Specifically, the "standard music sample" represented by the currently specified parameters
Is determined by, for example, the sample content determination unit 29 of the control center 20, and is downloaded from the control center 20, and the music content data is reproduced (output from the speaker 54). For this purpose, the sample content determination unit 29 uses a one-to-one
I am preparing sample music for the relationship. Then, one sample music corresponding to the value of the parameter designated according to the current position of the lever 61 is transmitted. Also, lever 6
The designated value of the parameter by 1 has a width. For example, sample music corresponding to the center value of the width can be determined.

Even if the sample music itself is not very similar to the music to be searched, when the searcher decides the parameter value through trial and error, only the adjective is used. This is useful because it is possible to predict to some extent how the search target music changes due to changes in individual parameter values that are difficult to understand.

In the present embodiment, a search / recommendation system for music content is taken as an example.
Assuming a recommendation system, a "standard image sample" represented by the currently specified parameters will be displayed.

Further, the allowable width setting buttons 62 displayed below the lever 61 are of a toggle type. By operating these with a pointing device, or by touching with a finger in the case of a touch panel type, the relevant parameter is set. Can be expanded or narrowed. For example, even if the value of each parameter has a value of seven steps across the center value “0” in the range of “−3” to “+3”, if the number of parameters is 6, the control center 20 searches. Combinations of the set values of the parameter set 16 transmitted as conditions include 7 to the sixth power,
That is, there are 117,649 combinations.

Although it depends on the scale of the prepared sound source database, if all parameters are set too fine, there is a high possibility that the corresponding work does not exist. Therefore, by setting the permissible range from the central value in each parameter, such a situation can be avoided, and the number of recommended titles according to the convenience of the searcher can be obtained.

For example, normally, the tolerance is set to “1”, and the range up to “1” point above and below the specified value can be permitted. The permissible width setting button 62 is operated. As a result, the tolerance becomes “2”,
Allows to accept parameter values up to a point.

When the allowable width setting button 62 is operated, up to three points are allowed, and all values are allowed. When the allowable width setting button 62 is operated again, it is possible to return to the original setting of the allowable degree 1 by operating the allowable width 0 again and excluding the specified value.

Further, as shown in FIG. 6, the magnitude of the tolerance that changes each time the tolerance width setting button 62 is operated is displayed together with the description such as “wide”, “normal”, and “all” on the button itself. If the configuration is expressed as a difference in the shape (up and down width) of the setting lever 61, it is easy for the user to intuitively understand how the tolerance of the parameter is currently set.

Further, the currently pre-filtered genre is displayed in the search genre field 63 provided at the bottom of the parameter setting screen 60, and the user changes the setting in the work number display section 64. The number of applicable works that change each time is displayed. By doing so, the recommended number of titles is notified in advance according to the content of each designated parameter set, so that the task of narrowing down the parameters becomes easier.

Finally, when the user touches the “Set” button 65 on the right side at the bottom, a content group (work group) corresponding to the parameter set by the searcher is displayed on the recommended content list display section 56 in FIG. . In this case, the content which is relatively highly evaluated by the searcher in the corresponding work group is displayed according to a predetermined priority rule such as a top-left or a left-left (content-oriented recommendation).

Further, the above-mentioned magnetic card or the like is inserted into the slot 5
3 and the personal ID is notified (transmitted) to the control center 20, the content that the searcher has already listened to can be omitted from the recommended content. In addition, in this case, since the user behavior data management unit 24 analyzes the user behavior database 25 that collects records on the setting tendency of the search parameters of the individual searchers, the present searcher can use similar search patterns. If it is determined that the user belongs to a "specific attribute group" composed of people having the above, the "recommend our shop" button 55b arranged in the middle of the genre button group is operated, Are displayed on the recommended content list display section 56 (personal oriented recommendations).

Next, when the searcher performs an operation of designating the item of the content that is actually interested from the list of recommended contents in the recommended content list display section 56, FIG.
Is displayed, and the content can be previewed.

The content preview screen 70 of the present embodiment
The album jacket (the image held by the accompanying information 15) of the corresponding content is displayed on the image display section 71 on the upper left side of the screen, and the list of songs included in the album of the corresponding content (the accompanying information is held) is displayed on the tune display section 72. Character information) is displayed. Then, by selecting a song on the song display section 72 with a mouse or a keyboard and operating an operation button including a rewind button 73a, a stop button 73b, and the like, a rewind or stop operation of the song can be performed.

Further, if a parameter confirmation display section 74 for displaying the parameter set 16 of the content and its set value is provided below the image display section 71, the searcher can set the parameter set on the control center 20 side. This is convenient because each set value of the set 16 can be more easily recognized.

When the "view to menu" button 75 is operated after previewing the content, an evaluation screen 80 for the content is displayed as shown in FIG. 8 before the recommended content list display section 56 is displayed. . Evaluation screen 80
A content evaluation column 81 is provided in the upper part of the box, and it is possible to evaluate how much the user likes the searched music in ten stages. Also, the artist information request column 82
In, when the degree of satisfaction is extremely high, for example, whether or not the user is interested in a work related to the content is set.

At the center of the evaluation screen 80, a parameter evaluation display section 83 for the content is provided. In the initial state, the parameter evaluation display unit 83 displays the parameter setting values of the content stored in the content parameter database 26. When the user slides the content evaluation lever 84 of each parameter up and down, The parameter setting value can be changed (evaluated).

That is, at the same time that the user evaluates how much the user likes the content, the set value of the content parameter set on the control center 20 side is displayed. If you find that the settings are not valid, you can modify the data. These pieces of information are temporarily stored in the search terminal 33 as log files, but are transmitted to the parameter setting unit 21 and the user behavior data management unit 24 of the control center 20 after the business day.
It is used to modify existing setting parameters of the content in order to further improve the expected satisfaction of the recommended content.

<Information Processing Flow in Music Content Search / Recommendation System> The information processing flow of the music content search / recommendation system in the present embodiment will be described below, taking a storefront mirror database installation type as an example.

First, as shown in FIG. 9, since the present embodiment is of a storefront mirror database installation type, the control center 20 sends each content provided by the content data providing unit 10 every time a new music CD is released. Additional information 15, sample sound source data 14, parameter setting unit 21
Is distributed to the search terminal 33 of each store 31. These are stored in a predetermined specific directory of the storage device of the terminal, and form a mirror database 42 in which the content database 27 and the content parameter database 26 are combined.

Here, the mirror database 42 of each search terminal 33 does not need to mirror all information contained in the database in the control center 20. For example, if you want to focus on listening to new songs, it is enough to store information on content released in recent months. By doing so, it is possible to reduce the initial investment amount related to each search terminal 33 having a huge storage capacity.

Similarly to the above-mentioned database, the user behavior database 25 can be mirrored on the search terminal 33 side. However, the user behavior database 25 stores more significant recommended data as the number of target users increases. As the possibility of withdrawal increases, the control center 20
It is desirable that the user behavior database 25 be directly accessed. In this case, of course, the user behavior data management unit 24
Function is also performed in the control center 20. In the present embodiment, it is assumed that the user behavior database 25 and the user behavior data management unit 24 are in the control center 20.

FIG. 10 shows an example of the structure of such a mirror database record. In the mirror database record in the present embodiment, the music accompanying information (music name, artist name, etc.) transferred from the content database
And a parameter definition value of each content transcribed from the content parameter database.

The (6) -th item describes the address of the sample sound source data file of each song included in the content. The item (7) describes the image file address of the album jacket data.
If local storage of compressed audio data files and album jacket image files for preview is not permitted,
The file may be referred to via the communication line 2 (such as the Internet) based on the address. On the other hand, when local storage of the file is permitted, the search terminal 33 does not refer to the address of the file,
What is necessary is just to refer to the file in a predetermined local directory in the storage device attached to the search terminal 33. In the item (14), the parameter name of the content parameter is described as an array, and in the item (15), the corresponding value is described as an array.

It is desirable that the parameters of the content parameter set 16 be selected as universally as possible. However, if the required parameters change as the times change, the parameter names are changed. Will be done.

FIG. 11 is a flow chart of information processing related to an operation on the recommended content list display section 56 in the search terminal 33. After the above-described mirror data is distributed and stored in each search terminal 33 (step S001), the search terminal 33
Is started, the search terminal 33 displays the content names and jacket images of all the contents on the recommended content list display section 56, including those newly received and stored (step S002). The portion that cannot be displayed at a time may be displayed by the scroll bar 58 (shown in FIG. 6).

Next, step 100 is started. In step 100, first, the card slot 5 of the search terminal 33
It is confirmed whether or not a user ID card has been inserted in Step 3 (Step S101). If it has been inserted, the user ID is stored in the terminal memory 44 (step S10).
2) The user ID is also recorded in the logging file 45 that records the operation status of each operation by the user. If no ID card is inserted, I
D = 0 is recorded in the logging file 45 (step S103).

The logging file 45 is later downloaded to the control center 20 for analyzing the behavior of the service user. However, the control center 20 does not manage a specific personal profile or the like. By doing so, it is possible to prevent the personal information of the service user, that is, the store customer, from being unnecessarily spread.

Next, the search terminal 33 enters a mode for accepting an operation event by the searcher (step S104). When the searcher touches the genre setting button 55 (case 1), the selected genre is set and stored (step S105), the content of the currently displayed recommended content list display section 56 is temporarily hidden, and the designated genre is designated. Only the contents included in the genre are displayed again (step S106).

On the other hand, when the searcher touches the parameter setting screen display button 57 (case 2), the parameter setting screen 60 shown in FIG. 6 is displayed (step S1).
07). Then, the process proceeds to step 200.

FIG. 12 is an information processing flow (step 200) related to an operation on the parameter setting screen 60 in the search terminal 33. By the searcher's operation on the parameter setting screen 60 described above, the value of the parameter expressing the music that the searcher wants to listen to at present is set to the parameter setting lever 61 (case 11) and the parameter allowable width setting button 62 (case 12). ), The sample music corresponding to the setting parameter is played (step S202), and at the same time, the parameter setting status of each time is sent to the recommended content determination unit 28 (step S203). Move to 300. After the corresponding content is extracted by the recommended content determination unit 28 in step 300, the number of the content is fed back to the display portion of the parameter setting screen 60 of the search terminal 33 (step S204, step S20).
5).

FIG. 13 is an information processing flow showing step 300 in recommended content determination section 28. As shown in the figure, the feedback of the number of relevant contents based on the current search condition to the parameter setting unit 21 is:
It differs depending on whether the genre is specified and the personal ID is specified.

First, the number of contents having the set value of the designated parameter set 16 is extracted from the contents parameter database 26 (step S301). Next, it is determined whether or not a genre is designated (step S302). If the genre is specified, the genre described in the supplementary information file of each content stored in the search terminal 33 is extracted only when the genre matches the specified genre (step S303).

Next, it is determined whether or not a personal ID has been designated (step S304). If a personal ID has been designated, the user access to the user action database 25 (step S305). Based on the personal use information accumulated by a process described later, the content that the user has already listened to can be omitted (step S306). Then, the number of extracted contents is returned to the content parameter setting screen 60 of the search terminal 33 (step S307).

Returning to the flow shown in FIG. The searcher performs trial and error on the parameter setting screen 60 (case 11).
And the operation of case 12), a specific search condition is determined, and when the set button 65 is selected (case 13), the display of the parameter setting screen 60 ends (step S206), and the recommended content determination unit The process moves to the process flow (step 400) in 28.

FIG. 14 is a diagram showing a processing flow by the recommended content determining unit 28. When the searcher touches the set button 65 on the parameter setting screen 60, a list of IDs of the corresponding contents is created (step S401). Next, it is determined whether or not a genre is specified (step S402). As in the case of step 300, when a genre is specified, only contents corresponding to the genre are extracted (step S403).

Next, the process proceeds to step 500, where the user behavior data management unit 24 is accessed, and among the extracted content groups, the content with a relatively high evaluation from the searcher is displayed preferentially (content / content). Oriented recommendations).

FIG. 15 is a flowchart for determining the content to be preferentially introduced by the user behavior data management unit 24 in step 500. User behavior data management unit 24
Accesses the user behavior database 25 (step S501), and the content-specific evaluation information 17 input by each user in the display of the content preview screen 70 described later.
Inquiry about the average evaluation point value for each content calculated from (step S502).

Next, based on the information, the recommended content I
The D list is re-sorted from the one with the highest average evaluation point (step S503), and returned to the recommended content determination unit 28 (step S504).

Next, the processing returns to the processing flow of FIG. After the processing of step 500, it is determined whether or not the personal ID has been designated (step S404). When the personal ID has been notified to the control center 20, the user behavior database 25 is accessed (step S405). ),
As in the case of step 300, among the recommended contents described above, those that the searcher has already listened to are sampled (step S406). Step 6
00, the user behavior data management unit 24 executes the genre selection button 55 displayed on the screen of the search terminal 33.
Is searched for a content ID to be displayed on the "Recommended shop" genre button 55a.

FIG. 16 is a processing flow of the user behavior data management unit 24 showing step 600. Step 60
When the searcher enters “0”, the user behavior data management unit 24 analyzes the user behavior database 25 that has collected records of past search parameter setting trends of the searcher (step S601), and the searcher determines that a specific search has been performed. It is determined whether there is a high possibility of belonging to a “certain searcher group” with a large number of parameter settings (step S602).
If it is determined that there is a high possibility, n pieces of content with high evaluation in the attribute group are extracted from the user action database in the order of the average evaluation point regardless of the content parameter set value specified this time. (Step S603). The extracted number n may be set in advance by the system side, or may be designated by a searcher in a series of search processes.

Then, as in the case of step 300, of the above-mentioned recommended contents, those which the searcher has already listened to are excluded (step S604),
The contents are sorted in descending order of evaluation and transmitted to the recommended content determination unit 28.

Returning to the flow of FIG. The recommended content for each individual determined in the above process is displayed in the “recommended for our shop” button 55 a in the content list display unit 56.
Is registered (step S407). When this button is touched by the searcher, the corresponding content is displayed (step S408): (Personal Oriented Recommendation).

In the present embodiment, “Recommended by our shop”
Only when the button 55a is actively pressed by the searcher, the recommended content for each individual is displayed in the content list display section 5.
6, the contents may be displayed first when the ID card is inserted into the slot when the personal ID is confirmed by the ID card.

FIG. 17 is a processing flow (step S700) of the designated content preview screen 70 of FIG. In step S701, it is determined whether or not the touch panel is touched. When the user touches the song name in the song display section 72 displayed on the right side of the screen (case 21), the audio file of the song is played (step S70).
2). When the rewind button 73a is touched (Case 2
2) The reproduction of the currently reproduced file is stopped (step S703), and the music file is reproduced from the beginning (step S704). When the stop button 73b is touched (case 23), the reproduction of the music is stopped (step S705).

[0112] The stop button 7 for each song during the continuous playback mode
When 3c is touched (Case 24), the continuous playback mode is terminated (Step S706), and the display of the stop button for each song is switched to "Continuous playback" (Step S707). On the other hand, when the continuous playback button 73c is touched in the song-by-song stop mode (case 24), the song-by-song stop mode is terminated (step S).
706), the display of the continuous playback button 73c is set to "stop for each song".
(Step S707).

If the search terminal 33 is connected to the store inventory management system online, an inventory confirmation button 76 is displayed on the screen, and when the button 76 is touched (case 25), the inventory confirmation button 76 is displayed. (Step S708), and the result (stock quantity) is displayed on the screen (step S709).

When the user finally touches the menu button 75 (case 26), the process proceeds to step 800, where the content evaluation screen 80 shown in FIG. 8 is displayed (case 2).
6).

FIG. 18 is an information processing flow in the content preview screen. Content evaluation column 8 at the top of the screen
1 (case 31), when the artist information desired field 82 (case 32) is touched, the specified information is temporarily stored in the memory 44 of the search terminal 33 (steps S801 and S802).

When the current content parameter value is changed by touching the content evaluation lever 84 (case 33) at the center of the screen, the information is used as change request information of the existing content parameter by the searcher.
It is stored in the memory 44 of the search terminal 33 (step S8)
03). Then, when the END button 85 at the bottom of the screen is touched (case 34), the above-mentioned memory storage information is transmitted to the user behavior data management unit 24 (step S80).
4).

FIG. 19 is an example of a record structure of the user behavior database 25 registered by the user behavior data management unit 24. Each record contains the ID of the searcher, the ID of the previewed content, the rating score given to the previewed content, the presence / absence of provision of artist information of the content (if a personal ID is announced),
In addition to the presence / absence of genre designation, information related to parameter search is described.

Each record is temporarily stored in the search terminal 33 every time the user of each store 31 uses the search terminal 33, distributed to the control center 20 at regular intervals, and finally stored in the all search terminals 33. The control center 20 as a user behavior database 25 covering information
Held within.

The user behavior data management unit 24 feeds back corresponding information from the database to the search terminal 33 according to the search request condition from each search terminal 33. When a parameter change request has been made by the user, the information is transferred to the parameter setting unit 21, and the parameter setting value of each content is corrected by the parameter setting unit 21 if necessary.

Further, if the information on the parameter setting status of the searcher described in this record is analyzed, it can be seen that despite the frequency of setting by the searcher being relatively high, relatively few contents are supplied. The parameter set 16 can be derived. This information is
Since it is positioned as information related to the type of content area that the user is potentially seeking but does not provide enough content, use it as marketing information for content creators in the future of content creation Becomes possible.

In this example, the search terminal 33 at the store is used.
Although the present invention has been described using an example, the personal computer (home) 32 connected to the general Internet
Alternatively, a mobile phone or a mobile terminal having a liquid crystal display function can be used as a search terminal.

Further, in the present embodiment, the present invention has been described with a music listening service as a specific example. However, the present invention is used as a channel selection interface of a music specialty station in a digital multi-channel radio receiving terminal in which the number of channels is expected to increase rapidly in the future. It is also possible to use. In this case, if a conventional radio channel selection interface is used, the user can actually listen to the broadcast content of the selected station, or broadcast his favorite music if he has not obtained detailed program information on the music broadcast at the station in advance. Although it was not possible to select the radio station that is performing, using the present invention, each music broadcasting station simultaneously broadcasts the parameters of the music currently being broadcast in a data broadcast or the like with the main broadcast, Since the listener can select a broadcasting station with a simple interface and listen to a favorite music, it is possible to reduce the labor and time involved in selecting the station.

[0123]

The content search / recommendation system according to the present invention has the following effects in retrieving and recommending unknown contents, such as new music, which many of the searchers have not yet enjoyed.

According to the first aspect of the present invention, in the content search / recommendation system, since the control center includes the content parameter database that expresses the characteristics of each content by a combination of a plurality of numerical parameters, Even unknown contents can be easily searched for by specifying the attribute of the desired contents by the contents parameter.

For example, even if the searcher does not know the music name of the searched music, the name of the player, etc., the user can search for the music by designating the type of music he / she wants to hear at each time through the content parameter. Is possible, so that a wider range of music can be sampled.

Further, even when it is difficult to immediately acquire the preference correlation data based on the evaluation data of a considerable number of searchers, using the parameter set hypothetically defined on the control center side provides a sufficiently effective method. Recommended content can be provided to searchers from the beginning.

According to the second aspect of the present invention, a content search / recommendation system includes a search terminal having a lever for variably specifying each parameter value and allowable range of the content parameter database. , The attributes of the content to be searched can be easily specified and fine-tuned, so that the search operation is made more efficient.

Further, according to the invention set forth in claims 3 and 4, the search terminal according to the present invention is adapted so that each time a combination of numerical values of each content parameter is specified by the user, the search terminal responds to the specified parameter. Since there is a means for previewing the sample content to be performed, it is possible to directly recognize the influence of the parameter designated by itself on the search content, and thus the search operation is made more efficient.

Therefore, the searcher can gradually specify the attribute of the search content by using the parameter lever or the like, and the type of music that changes according to the parameter can be confirmed each time by listening to the representative music or the sample music. So, rather than just searching for adjectives,
It is possible to search more efficiently.

According to the fifth aspect of the present invention, after each user uses a content at the search terminal, parameter modification request information performed on each content is collected through a network line, and parameter change is performed. The parameter setting section that performs the parameter setting makes it possible for the searcher to constantly verify the parameter set of the hypothetically defined content on the distribution side, so that more objective content can be defined over time. become.

According to the invention described in claim 6, a user behavior data management unit that collects, via a network line, a content parameter specified by each user before using the content at the search terminal, A user behavior database that stores information, and a recommended content determination unit that determines recommended content for each user from the specification information of the content parameters stored in the database. Without deciding on the factors, the reasons why searchers prefer individual content compared to systems that test for "resulting" significant correlation among a large number of searchers Since it can be determined by a combination of the values of the "content parameter set", the predicted behavior of the searcher Less likely to be scattered, it is possible to perform more significance highly effective content recommendation.

Furthermore, it is possible to find “latent demand content” in which the content is not actually provided, despite the high degree of user demand. That is,
Identify potential demands for content (areas) that searchers are potentially seeking but are less satisfied with,
It is possible to derive marketing information on future content production and supply.

According to the seventh aspect of the present invention, since the contents searched and used by the user in the past are omitted from the search results, the user can select the contents from the search results when the number of search results is large. The task of selecting desired content is facilitated. Further, the time required for downloading via the communication line is also reduced, so that the operability is improved.

[Brief description of the drawings]

FIG. 1 is a basic configuration diagram of a content search / recommendation system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an overall schematic configuration of a content search / recommendation system according to an embodiment of the present invention.

FIG. 3 is a diagram showing an example of a music content parameter set in the content search / recommendation system.

FIG. 4 is a diagram showing an example of a definition record of each song in a content parameter database.

FIG. 5 is a schematic view of a shop terminal of the music content search / recommendation system.

FIG. 6 is a diagram showing a content parameter setting screen of the music content search / recommendation system.

FIG. 7 is a diagram showing a content preview screen in the music content search / recommendation system.

FIG. 8 is a diagram showing a content evaluation screen in the music content search / recommendation system.

FIG. 9 is a configuration diagram of a store terminal provided with a mirror database.

FIG. 10 is a diagram showing an example of a record structure of a mirror database in the music content search / recommendation system.

FIG. 11 is a flowchart relating to an operation on a recommended content list display unit on a search terminal in the music content search / recommendation system.

FIG. 12 is a flowchart related to an operation on a parameter setting screen on a search terminal in the music content search / recommendation system.

FIG. 13 is a flowchart of the recommended content determination unit in step 300 shown in FIG.

FIG. 14 is a flowchart of a recommended content determination unit in step 400 shown in FIG.

FIG. 15 is a flowchart relating to determination of a preferential introduction content by a user action management unit in step 500 shown in FIG. 14;

FIG. 16 is a flowchart of a user action management unit in step 600 shown in FIG.

FIG. 17 is a flowchart of a designated content preview screen on a search terminal in the music content search / recommendation system.

FIG. 18 is a flowchart when a button is pressed on a menu on a designated content evaluation screen on a search terminal in the music content search / recommendation system.

FIG. 19 is a diagram illustrating an example of a record structure of a user action database registered by a user action management unit.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Content data provision part 20 Control center 21 Parameter setting part 24 User behavior data management part 25 User behavior database 26 Content parameter database 27 Content database 28 Recommended content determination part 29 Sample content determination part 30 Service utilization part 31 Store 32 Individual user 33 Search terminal 34 Search terminal 51 Personal computer 52 Display with touch panel 60 Parameter setting screen 70 Content preview screen 1 Evaluation screen

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 17/60 172 G06F 17/60 172 302 302E 326 326 502 502 502 F term (Reference) 5B049 AA01 AA04 BB11 BB49 CC08 DD01 EE05 EE28 FF03 FF06 GG04 GG07 5B075 KK07 KK13 KK33 KK37 KK40 ND14 ND20 ND23 NK02 NK04 NK06 NK39 PP03 PP10 PP28 PQ04 PQ32 PQ42 PR03 PR10 5E501 AB13 AC25 AC33 CA02 CA06 FA05 EB02 EB02

Claims (7)

[Claims] 1. A system in which content stored in a control center is used by a search terminal connected to the control center via a communication line, the control center comprising: a content database for storing the content; A content parameter database that stores sensory characteristics of the content as a parameter set, which is a set of one or more quantified attributes, wherein the search terminal performs a search operation using the parameter database. A content search / recommendation system which receives a recommendation of a content group from the content database according to a search condition, and makes a desired content available from the content group. 2. The content search / recommendation system according to claim 1, wherein the search terminal displays a display and a parameter setting screen on which a search condition can be set for the parameter set. Means for displaying an operation lever capable of designating a search condition having an allowance for each attribute of the parameter set by sliding on the parameter setting screen; and performing the operation according to the allowance Means for visually displaying a designated width when a search condition is designated by a lever on the parameter setting screen. 3. The content search / recommendation system according to claim 1, wherein the control center further comprises sample content determination means for determining sample content corresponding to a set value of the parameter set. The content search / recommendation system further comprising: means for previewing the sample content corresponding to the search condition every time a search condition is set in the parameter set. 4. The content search / recommendation system according to claim 1, wherein the content database is a database for storing music content, and the control center is further configured to set the parameter set. Sample content determining means for determining a sample music corresponding to the value, wherein the search terminal further comprises means for reproducing the sample music corresponding to the search condition every time a search condition is set in the parameter set. Content search / recommendation system characterized by having 5. The content search / recommendation system according to claim 1, wherein the search terminal can further input a parameter correction request for each content after the user uses the content. The control center further collects, via the communication line, the parameter correction request information input on the evaluation screen, and changes the set value of the parameter set. A content search / recommendation system characterized by comprising parameter setting means for causing the content to be searched. 6. Claims 1 to 3 and 5
The content search / recommendation system according to any of the above, wherein the control center further collects a search process of the search terminal for each user identified by a predetermined determination unit. Management means; a user behavior database that stores information collected by the user behavior data management means; A content search / recommendation system comprising: a content determination unit. 7. The content search / recommendation system according to claim 6, wherein the recommended content determination unit further excludes, from the search result specified by the user, the content used by the user in the past. A content search / recommendation system characterized by comprising means.