JP3547339B2 - Preference information collection system - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a preference information collection system that efficiently collects individual preference information required for collaborative filtering.
[0002]
[Prior art]
2. Description of the Related Art Collaborative filtering is used as an information retrieval technique in an area where personal characteristics are difficult to extract by a computer, such as personal preference, which requires human subjective judgment. In this collaborative filtering, human preferences are grouped by calculating the similarity of the evaluation for each preference object from a preference profile indicating the subjective evaluation (degree of interest) of the human preference for the preference object. This is a technique for estimating the preference of an unevaluated person by judging that similar evaluation is performed for the same preference target.
Based on the presumption that it is likely that other users in the same group who are likely to receive a high rating among the unevaluated preference subjects will receive a higher rating, this technique is referred to as "Recommended". It can be presented to other users as "information."
[0003]
[Problems to be solved by the invention]
In the collaborative filtering, if the preference information is small, for example, if the number of preference profiles themselves or the number of preference objects serving as a material for creating the preference profile is small, it is determined that all persons belong to the same group, or conversely, all the persons belong to the same group. In some cases, it may be determined that a person belongs to a different group, and credible results cannot be obtained. Therefore, in order to make collaborative filtering more effective, it is necessary for many people to evaluate many preference objects and accumulate the results as preference information.
[0004]
However, if the number of preference objects is excessively increased, the clear reason and motivation for performing evaluation for each person are relatively reduced, and it becomes impossible to expect a correct evaluation result. Also, a preference target that is not originally set for collaborative filtering cannot be used as an evaluation target. Furthermore, when a preference object is newly added, it cannot be used as a preference profile unless it is evaluated by some people.
[0005]
Therefore, an object of the present invention is to provide a preference information collection system capable of efficiently collecting preference information of many individuals.
[0006]
[Means for Solving the Problems]
A preference information collection system according to the present invention that solves the above-mentioned problems includes: a presentation unit that presents a preference target to be a creation material of a preference profile in a place of electronic communication activity by a plurality of users; Evaluation information receiving means for receiving evaluation information input by the user who is interested in crab for the preference object; evaluation information storage means for storing the received evaluation information together with identification information of the user who has input the evaluation information; A preference profile management means for creating or updating the user's preference profile based on the accumulated evaluation information; and collecting information representing the user's personal preference from a communication activity.
[0007]
The presenting means, for example, presents a list of classification names of preference targets classified in advance into a plurality of groups, and when any one of the classification names is selected according to the presentation, the preference target belonging to the classification name is selected. This is presented together with the evaluation input field. When a document database that holds a plurality of electronic documents partially including a preference target is provided, the presenting unit inputs a specific electronic document selected from the document database in accordance with a user's request into a preference target and its evaluation information. It is configured to be presented to the user together with the column. In this case, the preference target area in the electronic document is presented to the user in a form different from the area other than the preference target.
[0008]
Another preference information collection system according to the present invention includes a preference object holding unit that holds a preference object serving as a material for creating a preference profile, and an electronic document that receives input of an electronic document from a place of electronic communication activity by a plurality of users. Receiving means, a document analyzing means for checking whether or not the received electronic document includes a preference target, and, if so, estimating evaluation information for the preference target based on a context in which the preference target appears; Evaluation information storage means for storing the obtained evaluation information together with the identification information of the user who has input the electronic document; preference profile management means for creating or updating the user's preference profile based on the stored evaluation information; And collects information representing personal preferences of users from places of communication activities.
[0009]
The document analysis unit includes, for example, a context database storing positive expression information and negative expression information, and a unit for specifying expression information appearing after a preference target appearing in the electronic document. Is configured to estimate the evaluation information depending on which of the positive expression information and the negative expression information in the context database belongs to. The frequency of appearance of the preference target in the electronic document may be detected, and the estimated evaluation information may be corrected according to the detection result.
[0010]
Preferably, a document database that stores electronic documents for which estimation information has already been estimated, and means for searching whether or not there is an electronic document including the preference target in the document database when the preference target is added are provided. A preference information collection system is provided. The document analysis unit is configured to re-estimate the evaluation information for the electronic document when there is an electronic document including the added preference target.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a configuration diagram of a preference information collection system to which the present invention is applied.
The preference information collection system 1 is realized by a computer device (or system) having a storage device 15 in addition to an input device such as a keyboard, an output device such as a display, and the like. , An input / output control unit 11, an authentication processing unit 12, a data management unit 13, and a document analysis unit 14, which are formed by reading and executing a user database 151 and a preference target database 152 in the storage device 15. , A preference profile database 153, a document database 154, and a context database 155.
[0012]
The user database 151 manages a user of the preference information collection system 1 (hereinafter, “user”), an identifier of the user, and an authentication key, and is referred to by the authentication processing unit 12. It has become.
The preference target database 152 manages preference targets (movie titles, director names, actor names, etc. for movies) according to the classification (movie, fashion, etc.), that is, evaluation items serving as preference profile creation data. .
The preference profile database 153 manages a user's preference profile or profile data (evaluation information and the like) in an updatable manner.
The document database 154 manages an electronic document created by a user together with information such as a creator name and a creation date and time.
The context database 155 stores positive expression information and negative expression information in the context of the electronic document, which are used when the document analysis unit 14 analyzes the electronic document.
[0013]
The databases 151 to 155 are stored in one storage device (the storage device 15 in the illustrated example), and are distributed to a plurality of storage devices (including a case where they are connected by a communication line or the like). May be stored.
[0014]
The input / output control unit 11 controls input / output of various information exchanged with a plurality of users (through a communication line, an input device, and an output device), and includes an input / output interface (not shown) and the like. Be composed. The authentication processing unit 12 performs an authentication process based on a user identifier and an authentication key input through the input / output control unit 11.
[0015]
The data management unit 13 comprehensively manages data processing performed by the system. Specifically, it performs processing for presenting a preference target to each user, processing for creating or updating a preference profile of the user based on evaluation information input from the user in accordance with the presentation, and the like. . The document analysis unit 14 performs an electronic document analysis process. For example, with respect to the electronic document presented in the above-described communication place, the part of the preference target is presented in a different form from the part other than the preference target by referring to the preference target database 152 and the context database 154 in the storage device 15, Analyzes the context of new documents created by users.
[0016]
Next, an operation mode of the preference information collection system 1 configured as described above will be described.
In the preference information collection system 1 of the present embodiment, a place for communication activities by a plurality of authenticated users, for example, an Internet website is provided. On this website, a user can freely create an electronic document, and the created electronic document is sequentially collected and stored in the document database 154. Other users can freely refer to and attach the related electronic document to the electronic document, and the creator of the original electronic document can create a response document to the related electronic document, Continuous exchange of questions, answers, discussions, etc. can be performed. When it comes to movies, you can exchange impressions about the work, and ask and answer questions about the performers and shooting locations. Each electronic document and its creator can be specified by following the reference relation.
[0017]
After confirming the authentication, the user selects to refer to an existing electronic document or create a new document. The processing procedure in the preference information collection system 1 in this case is as shown in FIG.
Referring to FIG. 2, first, the authentication processing unit 12 instructs the user to input a user identifier and an authentication key through an output device (step S201). When a pair of an identifier and an authentication key is input from the user, the authentication processing unit 12 refers to the user database 151 and determines whether the input pair of the identifier and the authentication key is valid (Step S202: Yes). , S203). If it is not correct, an instruction to input another authentication key or the like is issued (step S204: No). If it is valid, the processing authority is transferred to the data management unit 13.
[0018]
The data management unit 13 presents a selection screen of a category (movie, fashion, etc.) that interests the user through the input / output control unit 11 and refers to an existing document or creates a new document to provide a new topic. Then, the user is prompted to select whether to create a reply document for the existing document (step S204: Yes, S205).
If there is an input of the selection data from the user and it is a reference to an existing document, the process is performed according to the procedure shown in FIG. 3 (step S206: refer to the existing document). On the other hand, if it is the creation of a new document or the creation of a reply to an existing document (together, "document creation"), the process is performed according to the procedure of FIG.
[0019]
<Existing document reference processing>
3, the data management unit 13 causes the output device to present the reference electronic document stored in the document database 154 through the input / output control unit 11 (step S301). At this time, the preference target database 152 is searched to identify the preference target described in the electronic document for reference. Then, the color of the part of the electronic document that is the object of preference is changed so that the user can easily understand that the evaluation is possible. FIG. 5A shows a screen display example in this case. In the example of FIG. 5A, “AAA” 51, “BBB” 52, “CCC” 53, “DDD” 54, and “EEE” 55 in the electronic document 50 are preference targets, and these colors can be changed. , Or underlined.
[0020]
The user selects one of these preference targets using a pointing device such as a mouse. When the data management unit 13 detects that the preference target is selected by the user (step S302: Yes), the data management unit 13 instructs an evaluation input for the selected preference target (step S303). The display of the screen at this time is, for example, as shown in FIG. The illustrated example is an example in which the user selects “AAA” 51 in FIG. The user inputs evaluation information (numerical information or the like) into a predetermined place 511 of the evaluation input window 510.
[0021]
When the evaluation information is input (Step S304: Yes), the data management unit 13 updates the user's preference profile in the preference profile database 153 based on the evaluation information (Step S305). If the end instruction has been input, the process is terminated (step S306: Yes), and if not, the process returns to step S201 (step S306: No).
[0022]
<Process when creating documents>
Referring to FIG. 4, when document creation is selected in step S206, the data management unit 13 receives the electronic document through the input / output control unit 11 (step S401). Let them do things.
[0023]
First, the document analysis unit 14 refers to the preference target database 152 to identify what the preference target in the electronic document is (step S402), and determines the context in which the preference target appears. It is determined whether the evaluation is an evaluation or a negative evaluation, and the evaluation is determined (step S403). Specifically, the evaluation information is determined by referring to the context database 155 depending on whether a positive expression or a negative expression appears after the preference target. If necessary, the evaluation information is corrected according to the preference target or the appearance frequency of the positive expression / negative expression (step S404). When the same user creates an electronic document with a different evaluation for the same preference target, the electronic document with the latest creation date and time is adopted, and the evaluation information for the electronic document is used as the evaluation information of the user.
[0024]
The data management unit 13 updates the personal preference profile of the user in the preference profile database 153 based on the determined or corrected evaluation information (step S405). If there is another preference target to be used as the document, the processing after step S402 is repeated (step S406: Yes).
[0025]
When the analysis of all the preference objects is completed, the data management unit 13 stores the electronic document in the document database 154 (Step S407). If the end instruction has been input, the process is terminated (Step S408: Yes), and if not, the process returns to Step S201 (Step S408: No).
[0026]
When a preference target is added to the preference target database 152, there is a possibility that the added preference target has appeared in a document that has already been created. In this case, the document analysis unit 14 searches the document database 154 to check whether there is a document including the added preference target. If there is, the same processing as in steps S402 to S406 described above is performed, and the preference profile of the document creator is updated. As a result, even when a preference target is added, it can be easily reflected in the preference profile. Also, it is easy to estimate the evaluations of other users.
[0027]
As described above, according to the preference information collection system 1 of the present embodiment, it is possible to easily collect more preference information than in the conventional method, and to enhance the credibility and usefulness of the result of collaborative filtering. Can be.
[0028]
In the reference electronic document, it is also possible to evaluate all words as preference objects. In this case, there is an advantage that the management of the preference target database 152 becomes unnecessary.
[0029]
Further, it is also possible to classify the preference target into, for example, a movie, and to efficiently collect information by restricting topics when creating an electronic document.
[0030]
【The invention's effect】
As is apparent from the above description, according to the present invention, it is possible to obtain a unique effect that it is possible to extremely easily collect more preference information from a place of a normal communication activity.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of an information collection system to which the present invention has been applied.
FIG. 2 is a diagram showing a procedure of user authentication and the like of the information collection system.
FIG. 3 is a diagram showing a procedure of updating a personal preference profile when referring to an existing document.
FIG. 4 is a diagram showing a procedure of updating a personal preference profile when creating a document.
FIGS. 5A and 5B are diagrams showing display screens when an existing document is referred to.
[Explanation of symbols]
1 Preference information collection system 11 Input / output control unit 12 Authentication processing unit 13 Data management unit 14 Document analysis unit 15 Recording device 151 User database 152 Preference object database 153 Preference profile database 154 Document database 155 Context database 50 Electronic document display for reference Screens 51, 52, 53, 54, 55 Selectable preference objects 510 Evaluation input screen 520 Evaluation input window

Claims (9)

A preference information collection system in which accumulation of electronic documents by a plurality of users and the accumulated existing electronic documents can be referred to by other users ,
Presentation means for presenting a preference target that is included in the electronic document to be referred to and is a creation material of a preference profile,
Evaluation information receiving means for receiving evaluation information input for the preference object included in the electronic document referred to by the user who is interested in any of the presented preference objects,
Evaluation information storage means for storing the received evaluation information together with the identification information of the user who has input the evaluation information,
A preference profile management unit that creates and updates a preference profile of the user based on the accumulated evaluation information,
Preference information collection system characterized by collecting information indicating the preference of individual users. The presenting means presents a list of classification names of preference objects classified in advance into a plurality of groups, and evaluates the preference objects belonging to the classification name when any one of the classification names is selected according to the presentation. Characterized by being presented together with the input field,
The preference information collection system according to claim 1. A document database that holds a plurality of electronic documents including the preference target in part,
The presenting means is configured to present a specific electronic document selected from the document database in response to a request of a user to the user together with a preference target and input fields for evaluation information thereof. Do
The preference information collection system according to claim 1. The region of the preference target in the electronic document is presented to the user in a form different from the region other than the preference target,
The preference information collection system according to claim 3. A preference object holding means for holding a preference object serving as a material for creating a preference profile,
Electronic document receiving means for receiving input of an electronic document;
A document analysis unit that checks whether or not the preference target is included in the received electronic document, and if so, estimates document evaluation information for the preference target based on a context in which the preference target appears.
Evaluation information storage means for storing the estimated evaluation information together with the identification information of the user who has input the electronic document;
Preference profile management means for creating or updating the user's preference profile based on the accumulated evaluation information ,
The document analysis unit has a context database storing positive expression information and negative expression information, and a unit for specifying expression information that appears after a preference target that appears in the electronic document. Is configured to estimate the evaluation information depending on which of the positive expression information and the negative expression information in the context database belongs to,
Preference information collection system. The electronic document can be attached to a related electronic document by a user other than the creator, and the related electronic document can be referred to by a user other than the creator to associate the related electronic document. Each electronic document and its creator can be specified by tracing the reference relationship,
The preference information collection system according to claim 5. The document analysis unit detects the frequency of appearance of a preference target in an electronic document, and is configured to correct the estimated evaluation information according to the detection result.
The preference information collection system according to claim 5. A document database that stores electronic documents for which estimation information has already been estimated,
Means for retrieving whether or not an electronic document including the preference target exists in the document database when the preference target is added,
6. The preference according to claim 5, wherein the document analysis unit is configured to re-estimate the evaluation information for the electronic document when there is an electronic document including the added preference target. Information collection system. In the case of an electronic document having different evaluation information for the same user and the same preference target, the evaluation information based on the latest electronic document is configured to be the evaluation information of the user. A preference information collection system according to any one of claims 5 to 8 .

Images