JP7002883B2 - Evaluation device, evaluation method and evaluation program - Google Patents

Description

The present invention relates to an evaluation device, an evaluation method, and an evaluation program for evaluating an object.

Conventionally, an evaluation device for calculating an evaluation for an object such as a content or a restaurant is known (see, for example, Patent Document 1).
The evaluation device of Patent Document 1 gives reliability to each user, and calculates the evaluation of the object based on the reliability given to the user and the evaluation of the object by the user (user evaluation).

Japanese Unexamined Patent Publication No. 2011-192103

By the way, in the above-mentioned evaluation device, an object having a different category may be evaluated. In such a case, it is required to improve the accuracy of evaluation of the object.

An object of the present invention is to provide an evaluation device, an evaluation method, and an evaluation program capable of improving the accuracy of evaluation of an object.

The evaluation device of the present invention includes a weight value setting unit that sets a weight value of a user for each category of an object, and an evaluation calculation unit that calculates an evaluation of the object. The user and the user who evaluated the object from the user evaluation history data associated with the object, one or more users who evaluated the object, and a user evaluation which is an evaluation of the object by the user. It is characterized in that the evaluation is read out and the evaluation of the object is calculated based on the weight value corresponding to the object set for the user and the user evaluation.

In the present invention, the weight value setting unit sets the weight value for the user for each category of the object, and the evaluation calculation unit sets the weight value corresponding to the category of the object set for the user and the user evaluation. Based on the above, the evaluation of the object is calculated.
According to this configuration, by setting the weight value according to the reliability of the user for each category, the evaluation of the object can be calculated accurately even if the reliability of the user changes depending on the category of the object. , The accuracy of evaluation can be improved.

The figure which shows the schematic structure of the evaluation system in embodiment of this invention. The figure which shows the schematic structure of the evaluation apparatus in embodiment. The figure which shows an example of the user evaluation history data in an embodiment. The flowchart which shows the evaluation process in an embodiment. The figure which represented the reference user data (other user data) in an embodiment by a bar graph. The figure which shows the relationship between the reference user data and other user data in an embodiment.

[Embodiment]
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[Evaluation system configuration]
FIG. 1 is a diagram showing a schematic configuration of the evaluation system 1 of the present embodiment.
In the evaluation system 1, a plurality of user terminals 2 and an evaluation device 3 are connected so as to be able to communicate with each other via a network such as the Internet.
The user of each user terminal 2 operates the user terminal 2 to transmit the evaluation (user evaluation) performed on the object to the evaluation device 3. The evaluation device 3 accumulates user evaluations transmitted from each user terminal 2. Further, the evaluation device 3 sets the weight value of the user for each category of the object. Then, the evaluation device 3 calculates the evaluation of the object based on the weight value corresponding to the object set by the user and the user evaluation for the object.
In this embodiment, an example in which the object is a restaurant will be described.
Hereinafter, each configuration of the evaluation system 1 will be described in detail.

[User terminal configuration]
The user terminal 2 is a device such as a computer associated with the user. Examples of the computer include a portable terminal such as a smartphone and a tablet, a fixed terminal such as a personal computer (PC), and the like.
When the user inputs the store to be evaluated (store ID) and the evaluation of the store (user evaluation) by operating the user terminal 2, for example, when using a restaurant, the user terminal 2 is the user. The evaluation data including the (user ID) and the input store and user evaluation is transmitted to the evaluation device 3 via the network. The user evaluation is related to, for example, the taste of food, the atmosphere of the store, the attitude of the clerk, and the like, and is input on a five-point scale of, for example, "1" to "5".

[Evaluator configuration]
FIG. 2 is a block diagram showing a schematic configuration of the evaluation device 3.
The evaluation device 3 is composed of a computer and is connected to a network. The evaluation device 3 includes a communication unit 31, a storage unit 32, a control unit 33, and the like that enable communication with each user terminal 2.

[Structure of storage unit]
The storage unit 32 is a data storage device composed of, for example, a memory and a hard disk.
Various programs for controlling the evaluation device 3 are stored in the storage unit 32. Further, the storage unit 32 includes a history data storage unit 321.

FIG. 3 is a diagram showing an example of the data structure of the user evaluation history data 321A stored in the history data storage unit 321.
The user evaluation history data 321A includes a store (store ID), a user who evaluated the store (user ID), and an evaluation of the store by the user (user) for each restaurant category such as a restaurant, ramen shop, and cafe. Evaluation) are associated with each other and stored.

In the example of FIG. 3, for example, the restaurant category is associated with the stores S101, S102, S103 ..., the ramen shop category is associated with the stores S201, S202 ..., And the cafe category is associated with the stores S301, S302 ... It is associated.
Then, the user who evaluated the target store and the evaluation (user evaluation) of the target store by the user are associated with each of the stores S101, S102, S103 ..., S201, S202 ..., S301, S302 ... ing.
For example, in the restaurant store S101, the user U001, U002, U005, U007 ... Who evaluated the store S101 and the user U001, U002, U005, U007 ... evaluated the store S101 (5, 1, 3, 4 ...). Is associated with.

Further, in the example of FIG. 3, for example, the user U001 evaluates the restaurant stores S101 and S102, the ramen shop store S201, and the cafe store S301, and each of these stores S101, S102, S201, and S301. Is associated with user evaluation by user U001 and user U001.
Each data stored in these user evaluation history data 321A is acquired from the user terminal 2. If the same store is evaluated by the same user multiple times, only the latest data is stored.

[Control unit configuration]
Returning to FIG. 2, the control unit 33 is composed of a calculation circuit such as a CPU (Central Processing Unit) and a storage circuit such as a RAM (Random Access Memory), and provides an evaluation program (software) stored in the storage unit 32 or the like. By executing this, it functions as an evaluation data acquisition unit 331, a reference setting unit 332, a weight value setting unit 333, an evaluation calculation unit 334, and a fluctuation determination unit 335.

The evaluation data acquisition unit 331 acquires the evaluation data transmitted from the user terminal 2 via the communication unit 31, and stores the acquired evaluation data in the user evaluation history data 321A.
The reference setting unit 332 sets a reference user as a reference for evaluation.
The weight value setting unit 333 sets the weight value for each user for each category of the store. In the present embodiment, the weight value setting unit 333 sets the weight value when the reference user is used as a reference.
The evaluation calculation unit 334 calculates the evaluation of the store to be evaluated based on the weight value corresponding to the store to be evaluated set by the user and the user's evaluation of the store by the user.
The fluctuation determination unit 335 determines fluctuations in a plurality of user evaluations for each store.
Details of each of these functional units will be described in the following description of the evaluation process and the fluctuation determination process.

[Evaluation processing]
FIG. 4 is a flowchart showing an evaluation process executed by the evaluation device 3.
When the evaluation device 3 receives, for example, an evaluation request signal requesting the evaluation of a store in a predetermined category from the user terminal 2, the evaluation device 3 executes the evaluation process.
When the evaluation process is executed, the reference setting unit 332 sets the user who is the source of the evaluation request signal as the reference user (step S1).

Next, the weight value setting unit 333 reads out from the user evaluation history data 321A the store evaluated by the reference user in the predetermined category and the user evaluation of the store by the reference user as the reference user data (step S2).
For example, in the example of FIG. 3, when the predetermined category is a restaurant, among the stores S101, S102, S103 ... Associated with the restaurant, all the stores evaluated by the reference user in the past and the user evaluations of the store by the reference user are all. read out.

Further, the weight value setting unit 333 identifies a user (other user) other than the reference user among the users who have evaluated the store of the predetermined category in the past from the user evaluation history data 321A. Then, for each specified other user, the store evaluated by the other user in the past in the predetermined category and the user evaluation of the store by the other user are read out as other user data (step S3). In addition, step S2 may be executed after step S3.

Then, the weight value setting unit 333 compares the reference user data with each other user data, and calculates the similarity of the user evaluation (step S4).
Specifically, in this embodiment, the degree of similarity is calculated based on the tendency of user evaluation.
For example, the reference user data and each other user data can be represented by a bar graph in which the horizontal axis is the store and the vertical axis is the user evaluation, respectively, as shown in FIG.
In this case, it can be determined that the closer the pattern shape of the bar graph is, the more similar the tendency of user evaluation is.

That is, for example, in the example of the reference user data D1 and the other user data D2 shown in FIG. 6, the value of the user evaluation itself is not close. For example, in the reference user data D1, the user evaluation for the store S101 is "4" and the user evaluation for the store S102 is "5", whereas in the other user data D2, the user evaluation for the store S101 is "2". The user evaluation for the store S102 is "3". That is, the reference user has a higher user evaluation value than the other users.
However, the reference user data D1 and the other user data D2 have similar pattern shapes when represented by a bar graph. That is, the user evaluation of the stores S102 and S107 is relatively high, the user evaluation of the store S103 is relatively low, and the user evaluation of the stores S101, S105 and S106 is the user evaluation of the stores S102 and S107 and the user evaluation of the store S103. The value between. Therefore, it can be determined that the reference user data D1 and the other user data D2 have similar user evaluation tendencies.

In the example of FIG. 6, there is no user evaluation for the store S104 in the reference user data D1. In this case, for example, the pattern shapes are compared except for the user evaluation corresponding to this store S104.
Further, in the example of FIG. 6, there is no user evaluation for the store S108 in the other user data D2. In this case, for example, the user evaluation corresponding to this store S108 is set to, for example, a preset value, and then the pattern shapes are compared. The value is, for example, the average value or the median value of all user evaluations for the store S108.

The reference user data and each other user data may be represented by, for example, a line graph other than the bar graph. That is, any graph may be used as long as it has a pattern shape.

Returning to FIG. 4, next, the weight value setting unit 333 sets the weight value corresponding to the predetermined category for each other user based on the calculated similarity (step S5).
The weight value is set in increments of 0.1, for example, with the minimum value being "0" and the maximum value being "1". That is, the higher the similarity of user evaluations, the larger the weight value is set.
Here, another user whose user evaluation is similar to that of the reference user can be determined to be a reliable user. In other words, when the standard user evaluates based on a certain standard, the other user who also evaluates based on a certain standard is the other user who evaluates based on a different standard each time (evaluation). There is a high possibility that user evaluations will be similar to those of other users who are sloppy. Therefore, it can be determined that the higher the similarity of user evaluations, the higher the reliability. Therefore, by setting the weight value based on the similarity of the user evaluation, the weight value can be set according to the reliability of the user.
A preset weight value is set for the reference user. The weight value may be appropriately set, for example, depending on how much the user evaluation of the reference user is to be reflected in the evaluation calculated by the evaluation calculation unit 334.

Then, the evaluation calculation unit 334 calculates the evaluation of each store in the predetermined category based on the weight value corresponding to the predetermined category set by the user and the user evaluation by the user (step S6).
Specifically, the evaluation calculation unit 334 is set in the user evaluation history data 321A for each user evaluation associated with the store to be evaluated, with the user evaluation value set to the user associated with the user evaluation. Multiply the weighted value to calculate the weighted user evaluation.
That is, for example, when the user evaluation is "3" and the weight value set for the user associated with the user evaluation is "0.1", the evaluation calculation unit 334 sets the weighted user evaluation to 0.3. It is calculated as (= 3 × 0.1).
Then, the evaluation calculation unit 334 calculates a total value obtained by adding all the weighted user evaluations calculated for each user evaluation. Then, the value obtained by dividing the total value by the number of weighted user evaluations is used as the evaluation of the store to be evaluated.

For example, the user U001, U002, U004, U008 evaluates the store to be evaluated, and the user evaluations given by each user U001, U002, U004, U008 to the store are 3,2,5. It is assumed that it was 4. Further, it is assumed that the weight values of the categories corresponding to the stores set in each user 001, U002, U004, U008 are 0.1, 0.8, 0.3, 0.1.
In this case, the evaluation of the store is calculated as (3 × 0.1 + 2 × 0.8 + 5 × 0.3 + 4 × 0.1) ÷ 4 = 0.95.

Then, the evaluation calculation unit 334 transmits the evaluation calculated for each store in the predetermined category and the store to the user terminal 2 from which the evaluation request signal is transmitted via the communication unit 31 (step S7). Then, the control unit 33 ends the evaluation process.

[Variation judgment processing]
Next, the variation determination process executed by the variation determination unit 335 will be described.
The fluctuation determination unit 335 determines in the user evaluation history data 321A whether or not the fluctuation value in the predetermined period of user evaluation for each store is equal to or higher than a preset threshold value.
The fluctuation value is, for example, a fluctuation value in a predetermined period of the average value of all user evaluations for the store. The predetermined period is, for example, one hour, one day, one month, or the like.
Then, when the fluctuation value becomes the threshold value or more, the fluctuation determination unit 335 transmits the store where the fluctuation value becomes the threshold value or more to the user terminal 2 via the communication unit 31.
The fluctuation determination process may be performed at a timing different from the above evaluation process, or may be performed at the same timing as the evaluation process.

[Action and effect of the embodiment]
User confidence may vary depending on the store category. For example, a user with high reliability in the restaurant category may have low reliability in the ramen shop category.
According to the present embodiment, the weight value setting unit 333 can set the weight value according to the reliability for the user for each category of the store, and the evaluation calculation unit sets the evaluation target category set for the user. The evaluation of the evaluation target is calculated based on the corresponding weight value and the user evaluation.
Therefore, even when the reliability of the user changes depending on the category of the store as described above, the evaluation of the evaluation target can be calculated accurately and the accuracy of the evaluation can be improved.

In the present embodiment, the reference setting unit 332 sets the reference user, and the weight value setting unit 333 sets the weight value for the other user based on the similarity between the reference user data and each other user data.
According to this configuration, the weight values of other users whose user evaluations are similar to those of the reference user can be increased, so that the evaluation matching the hobbies and tastes of the reference user can be calculated.

In the present embodiment, the weight value setting unit 333 calculates the degree of similarity between the reference user data and each other user data based on the tendency of user evaluation.
According to this configuration, even if the value of the standard user and the user evaluation are different, the weight value of another user whose evaluation standard is similar to that of the standard user can be increased, and the user evaluation of the other user is evaluated by the evaluation calculation unit 334. It can be strongly reflected in the calculated evaluation.

In the present embodiment, the fluctuation determination unit 335 determines whether or not the fluctuation value in the predetermined period of user evaluation for each store is equal to or higher than a preset threshold value, and if the fluctuation value is equal to or higher than the threshold value, the fluctuation value is changed. The store that exceeds the threshold value is transmitted to the user terminal 2 via the communication unit 31.
According to this configuration, for example, when a change occurs in a certain store and the user evaluations by a plurality of users change all at once, the user can be notified of the store.

[Modification example]
It should be noted that the present invention is not limited to the above-described embodiment, but also includes the modifications shown below to the extent that the object of the present invention can be achieved.

[Modification 1]
In the above embodiment, the reference setting unit 332 sets the user who is the source of the evaluation request signal as the reference user, but the present invention is not limited to this.
For example, any user selected by the source user may be set as the reference user. In this case, for example, even if the user evaluation of the user of the transmission source is not stored in the user evaluation history data 321A, another user whose user evaluation is stored in the user evaluation history data 321A can be selected. The store evaluation can be calculated using the other user as the reference user.

Further, the reference setting unit 332 may set a plurality of users as reference users. In this case, the weight value setting unit 333 sets the weight value of another user based on the total value (or average value) of the similarity calculated for each reference user. According to this configuration, it is possible to calculate an evaluation that matches the hobbies and tastes of a plurality of users set as reference users.

[Modification 2]
Further, the reference setting unit 332 may set a reference pattern indicating a reference value of user evaluation for each store without setting a reference user. In this case, the weight value setting unit 333 compares the pattern shape between the reference pattern and the user evaluation for each store of each user, and calculates the degree of similarity. Then, the weight value setting unit 333 sets the weight value for each user based on the similarity. The reference pattern may be preset or may be set by the user.

[Modification 3]
In the above embodiment, the weight value setting unit 333 calculates the similarity of the user evaluation between the reference user data and each other user data based on the pattern shape of each data, but is not limited to this.
For example, the similarity may be calculated based on the magnitude of the value of the user evaluation in addition to the pattern shape.
Alternatively, the similarity may be calculated based on, for example, the number of stores that match the user evaluation values, instead of the pattern shape. Alternatively, the reference user data and other user data may be grouped by a statistical method using the Dirichlet process, and the similarity may be calculated based on the result of the grouping.

[Modification 4]
In the above embodiment, the weight value setting unit 333 sets the weight value of the user based on the similarity of the user evaluation between the reference user and each other user, but the weight value setting unit is not limited to this.
For example, when the reliability of each user is known in advance by a questionnaire or the like, a weight value may be set for each user based on the reliability.
Alternatively, the reliability of each user may be determined based on the commonality of user evaluations among the users, and a weight value may be set for each user based on the reliability. That is, it is judged that the higher the degree of commonality of user evaluation with other users is, the higher the reliability is, and the value of the weight value is also increased.

[Modification 5]
In the above embodiment, the evaluation device 3 receives an evaluation request signal requesting evaluation of a store in a predetermined category from the user terminal 2, executes an evaluation process, and then sends the user terminal 2 to each of the predetermined categories. We are sending store ratings, but not limited to this.
For example, the evaluation device 3 may receive an evaluation request signal requesting an evaluation of a specific store from the user terminal 2, execute an evaluation process, and then transmit the evaluation of the specific store to the user terminal 2. good.
Further, the evaluation device 3 may transmit the evaluation to another user terminal 2 selected by the user terminal 2 from which the evaluation request signal is transmitted, or to the user terminal 2 registered in advance in the evaluation device 3. For example, in the case of the above modification 1, the evaluation may be transmitted to the user terminal 2 of the user set as the reference user.

[Modification 6]
In the above embodiment, when the evaluation device 3 executes the evaluation process, the variation determination process is executed, and the information of the store whose variation value is determined to be equal to or higher than the threshold value in the variation determination process is added to the evaluation calculated by the evaluation process. It may be reflected.
For example, for a store whose fluctuation value is determined to be equal to or higher than the threshold value, it is unlikely that the evaluation of the current store can be appropriately calculated even if the evaluation is calculated based on the past user evaluation. Therefore, the evaluation calculated for the store may be corrected to be lower by, for example, a certain value. Alternatively, the evaluation calculated for the store may be invalidated.

[Modification 7]
In the above embodiment, the user evaluation history data 321A is stored in the storage unit 32 provided in the evaluation device 3, but the user evaluation history data 321A is not limited to this.
For example, the user evaluation history data 321A may be stored in a storage device provided so as to be connectable to a network, separately from the evaluation device 3.

[Modification 8]
In the above embodiment, the category is set for each type of restaurant, but the category is not limited to this. For example, a category may be set for each dish offered at the store. Alternatively, a category may be set for each location of the store.

[Modification 9]
In the above embodiment, an example in which the object is a restaurant has been described, but the object may be content such as a moving image, a hotel guest room, a travel tour, or the like. When the object is a moving object, the weight value setting unit 333 sets the weight value of the user for each category such as Japanese movies, Western movies, dramas, and sports. Further, when the object is a hotel guest room or a tour (travel), a user weight value is set for each category of the object such as domestic, overseas, and business.
Further, the evaluation device 3 can also be used for human evaluation such as personnel evaluation. In this case, the object is a person such as an employee. The weight value setting unit 333 sets the weight value of the user for each category of the object such as a general employee and an executive employee.
Further, the evaluation device 3 can also be used for route search of automobiles, trains, and the like. In this case, the object is a candidate route obtained by the search. The weight value setting unit 333 sets the weight value of the user for each category of the object such as a region.

1 ... Evaluation system, 2 ... User terminal, 3 ... Evaluation device, 31 ... Communication unit, 32 ... Storage unit, 33 ... Control unit, 321 ... History data storage unit, 321A ... User evaluation history data, 331 ... Evaluation data acquisition unit 332 ... Reference setting unit, 333 ... Weight value setting unit, 334 ... Evaluation calculation unit, 335 ... Fluctuation determination unit.

Claims (5)

A weight value setting unit that sets the user's weight value for each object category,
An evaluation calculation unit that calculates the evaluation of the object,
A reference setting unit for setting a reference user as a reference for evaluation of the object is provided.
The weight value setting unit is
Each of the objects and the object from user evaluation history data associated with a plurality of the objects, one or more users who have evaluated each of the objects, and a user evaluation which is an evaluation of the object by the users. A graph showing the relationship between the object and the user evaluation is generated for each user as the tendency of the user evaluation, and based on the similarity of the pattern shape of the graph, the user evaluation of the reference user and other than the reference user. The degree of similarity of the user to the user evaluation is calculated.
Based on the calculated similarity, the weight value corresponding to the object is set for a user other than the reference user.
The evaluation calculation unit
From the user evaluation history data, the user who evaluated the object and the user evaluation are read out.
An evaluation device for calculating the evaluation of the object based on the weight value corresponding to the object set for the user and the user evaluation. In the evaluation device according to claim 1,
The user evaluation history data includes the user evaluation for a plurality of the objects.
The weight value setting unit is
An evaluation device for calculating the similarity based on the tendency of the user evaluation for a plurality of the objects. In the evaluation device according to claim 1 or 2.
An evaluation device including a fluctuation determination unit for determining whether or not a plurality of fluctuation values of a plurality of user evaluations for an object in a predetermined period are equal to or higher than a preset threshold value. It is an evaluation method that allows a computer to evaluate an object.
The computer
A weight value setting step for setting a user weight value for each of the object categories, and
An evaluation calculation step for calculating the evaluation of the object and
A standard setting unit that sets a standard user that serves as a standard for evaluation of the object is carried out.
In the weight value setting step,
Each of the objects and the object from user evaluation history data associated with a plurality of the objects, one or more users who have evaluated each of the objects, and a user evaluation which is an evaluation of the object by the users. A graph showing the relationship between the object and the user evaluation is generated for each user as the tendency of the user evaluation, and based on the similarity of the pattern shape of the graph, the user evaluation of the reference user and other than the reference user. The degree of similarity of the user to the user evaluation is calculated.
Based on the calculated similarity, the weight value corresponding to the object is set for a user other than the reference user.
In the evaluation calculation step,
From the user evaluation history data, the user who evaluated the object and the user evaluation are read out.
An evaluation method characterized by calculating the evaluation of the object based on the weight value corresponding to the object set for the user and the user evaluation. An evaluation program that is loaded and executed by a computer.
An evaluation program comprising the computer functioning as the evaluation device according to any one of claims 1 to 3.

Images