Information processing device, information processing method, and information processing program
The information processing device improves user characteristic evaluations by calculating weighted scores from self-assessment and peer-feedback, addressing inaccuracies in existing systems by adjusting weights based on characteristic types and user behavior, resulting in more accurate assessments.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2023-04-14
- Publication Date
- 2026-07-29
AI Technical Summary
Existing evaluation systems, such as those described in Patent Document 1, do not accurately consider the variability in self-assessment and the reliability of other users' evaluations, leading to inaccurate assessments of user characteristics due to uniform treatment of subjective and objective evaluations, overestimation or underestimation of certain evaluation items, and the potential influence of malicious users.
An information processing device that calculates a user evaluation score by applying weighting coefficients to both self-evaluation and peer-evaluation scores, using a combination of rule-based and machine learning methods to adjust these weights based on the type of characteristic item, and incorporates user behavior and feedback to improve accuracy.
The system enhances the accuracy of user characteristic evaluations by appropriately handling over- or under-evaluated items, considering the variability and reliability of evaluations, thereby providing a more precise assessment of user skills and traits.
Smart Images
Figure 0007896538000001 
Figure 0007896538000002 
Figure 0007896538000003
Abstract
Description
Technical Field
[0001] The present invention relates to an information processing apparatus, an information processing method, and an information processing program for evaluating user characteristic items.
Background Art
[0002] In social networking services (SNS), a vast amount of data is accumulated daily through user input and the like. The accumulated data includes multiple types of data such as basic information of the user input by the user himself / herself and information expressing the characteristics of the user (characteristic information), interaction information with other users, and the like. There is a possibility that these data contain content different from the truth (facts), and the reliability of the data is considered to have a great impact on the quality of the service. Therefore, the development of technologies for enhancing the reliability of data is underway.
[0003] For example, a system for matching the needs of employers and job seekers, and a system for evaluating the sales ability in an organization including salespersons and managers who manage those salespersons by a plurality of common evaluation items independent of the organizational hierarchy are known (see, for example, Patent Document 1).
[0004] In the sales ability evaluation apparatus of Patent Document 1, for each question of the questionnaire, each person inputs the question evaluation score that he / she has evaluated himself / herself, and the self-evaluation score is calculated for each evaluation item from the question evaluation score according to the correspondence relationship of the question content. The response evaluation score and the visual evaluation score scored from the positions of each layer according to the correspondence relationship of the question content are input, and the objective evaluation score is calculated for each evaluation item from the response evaluation score and the visual evaluation score. Further, in the sales ability evaluation apparatus of Patent Document 1, the average score of the self-evaluation score and the objective evaluation score is calculated for each evaluation item for each organizational hierarchy. And this average score is taken as the final score.
Prior Art Documents
Patent Documents
[0005]
Patent Document 1
[0006] It is conceivable to evaluate user characteristics using the data described above. For example, when evaluating user characteristics (such as English conversation or programming skills; hereinafter also referred to as "user characteristics"), using evaluation information from others in addition to self-reported user evaluation information would allow for a more accurate assessment of user characteristics. Furthermore, depending on the content of the user characteristics, there may be items where subjective evaluation should be given more weight.
[0007] However, in the sales force evaluation device described in Patent Document 1, the final score is the average of the self-evaluation score and the objective evaluation score for the evaluation indicator of user characteristics. As a result, subjective evaluations by users and objective evaluations by other users are treated uniformly. Therefore, there is a problem in that it does not accurately consider evaluation indicators (for example, national qualifications, etc.) that should be emphasized when evaluating user characteristics (in this case, sales force).
[0008] Furthermore, for evaluation items such as "favorite things," even though the user's self-assessment score should be prioritized, the evaluation score can change due to the intervention of other users' evaluations. This can lead to a problem where the accuracy of the user's final score decreases.
[0009] Furthermore, it does not take into account the variability in self-assessment, and for some evaluation items, there is a difference in evaluation scale between users who have a high self-assessment and users who are humble about their self-assessment, making it difficult to accurately assess user characteristics. In addition, it does not consider the reliability of other users who evaluate a given user, and if a user's characteristics are evaluated by a malicious user, it becomes impossible to accurately assess the user characteristics of the evaluated user.
[0010] This disclosure is made to solve these problems and aims to provide an information processing device, an information processing method, and an information processing program that can appropriately handle other users' evaluations of a user's characteristic items, and can correct characteristic items (evaluation items) that are overestimated or underestimated by users, thereby improving the accuracy of the evaluation of a user's characteristic items. [Means for solving the problem]
[0011] An information processing device relating to one aspect of this disclosure is: An information storage unit that stores information about the user and information about the actions of other users regarding the user's information, A data acquisition unit that acquires information on characteristic items from the information storage unit, including information on the user and information on the behavior of other users. A user characteristic score calculation unit calculates a user evaluation score for characteristic items based on information about the user and information about the behavior of other users regarding characteristic items. Equipped with, The user characteristics score calculation unit is: A self-assessment score calculation unit calculates a user's self-assessment score based on information about the user regarding characteristic items, A peer evaluation score calculation unit calculates peer evaluation scores based on information about other users' behavior regarding characteristic items, A final evaluation score calculation unit calculates a user evaluation score for a characteristic item by setting weighting coefficients for the self-evaluation score and the peer evaluation score according to the characteristic item, and calculating the user evaluation score for the characteristic item based on the set weighting coefficients, the self-evaluation score, and the peer evaluation score. It is an information processing device that includes [a specific component / function].
[0012] With the above configuration, the information processing device according to one aspect of this disclosure calculates a user evaluation score by applying weighting coefficients to the user's own self-evaluation score and the peer evaluation score by other users, respectively, thereby improving the accuracy of the evaluation of the user's characteristic items.
[0013] In an information processing apparatus according to one aspect of this disclosure, the information storage unit may store multiple characteristic items as characteristic items. In this case, the final evaluation score calculation unit may set weight coefficients based on rules or machine learning, depending on the type of characteristic item. By setting the weight coefficients flexibly in this way, the accuracy of the user's evaluation of characteristic items can be further improved.
[0014] In an information processing apparatus according to one aspect of this disclosure, the user characteristic score calculation unit may perform preprocessing on the text data acquired by the data acquisition unit using a word vector dictionary.
[0015] The information processing method relating to one aspect of this disclosure is: Information regarding the user and information regarding the behavior of other users, specifically information related to characteristic items, is obtained from the information storage unit. Based on information about the user regarding characteristic items, a user self-assessment score is calculated. Based on information about other users' behavior regarding characteristic items, a peer evaluation score by other users is calculated. Depending on the characteristic item, weighting coefficients are set for the self-assessment score and the peer assessment score, respectively. This is an information processing method that calculates a user evaluation score for a characteristic item based on set weighting coefficients, self-evaluation scores, and peer evaluation scores.
[0016] The information processing program disclosed herein is: The process involves obtaining information on characteristic items from the information storage unit, among information on the user and information on the behavior of other users. A process of calculating a self-evaluation score by a user based on information about the user regarding characteristic items, a process of calculating an other-user evaluation score by other users based on information about the actions of other users regarding characteristic items, a process of respectively setting weight coefficients regarding the self-evaluation score and the other-user evaluation score according to the characteristic items, a process of calculating a user evaluation score regarding the characteristic items based on the set weight coefficients, the self-evaluation score, and the other-user evaluation score, and an information processing program for causing an information processing apparatus to execute the above.
Effect of the Invention
[0017] According to the present disclosure, it is possible to appropriately handle the evaluations of other users regarding the characteristic items of a user, and to optimize the characteristic items (evaluation items) that are over- or under-evaluated by the user. Thereby, it is possible to provide an information processing apparatus, an information processing method, and an information processing program that can improve the accuracy in the evaluation of the characteristic items of the user.
Brief Description of the Drawings
[0018] [Figure 1] It is a block diagram showing the configuration of the information processing apparatus according to the embodiment. [Figure 2] It is an example of estimating the relevance of a reply considering the semantic category according to the embodiment. [Figure 3] It is an example of scoring for user attributes according to the embodiment. [Figure 4] It is a flowchart showing an example of the scoring process executed by the information processing apparatus shown in FIG. 1.
Mode for Carrying Out the Invention
[0019] [[ID=Embodiments of the present invention will be described below with reference to the drawings. However, the invention claimed is not limited to the following embodiments. Furthermore, not all of the configurations described in the embodiments are necessarily essential for solving the problem. For clarity of explanation, the following descriptions and drawings have been omitted and simplified as appropriate. In each drawing, the same elements are denoted by the same reference numerals, and redundant explanations have been omitted where necessary.
[0020] <Configuration of the information processing device> First, the configuration of the information processing device according to this embodiment will be described. Figure 1 is a block diagram showing the configuration of the information processing device 1 according to this embodiment. The configuration of the information processing device 1 according to this embodiment will be described with reference to Figure 1.
[0021] As shown in Figure 1, the information processing device 1 of this embodiment is a device for providing a service that processes user information of registered users, and comprises a data acquisition unit 10, a user characteristic score calculation unit 20, and an information storage unit 30. The information processing device 1 may be connected to servers 2A, 2B~2N for providing social networking services (SNS) via a network 3 such as the Internet. Each server 2A, 2B~2N is a server that provides services such as Twitter®, Instagram®, TikTok®, YouTube®, or Facebook®.
[0022] Each server 2A, 2B-2N is equipped with a user database 2A1, 2B1-2N1 for storing user information. User databases 2A1, 2B1-2N1 store user information for each user using each SNS. The information processing device 1 may be configured so that each user of the information processing device 1 can set whether or not to allow the information processing device 1 to access their user information stored in the user databases 2A1, 2B1-2N1 of each server 2A, 2B-2N. If a user allows access by the information processing device 1 (for example, a link to the user page of each SNS is provided in the user information of the information processing device 1), the information processing device 1 can directly obtain user information not only from the information storage unit 30 but also from each user database 2A1, 2B1-2N1 in the scoring process executed by the information processing device 1, as described later.
[0023] The information storage unit 30 stores information about the user and information about the actions of other users regarding the user. As shown in Figure 1, the information storage unit 30 includes a user information database 31, a word vector dictionary database 32, and a weight information database 33.
[0024] The user information database 31 stores information about users, such as user attribute information, user self-assessment information, user behavior history information, and user feedback information. User attribute information includes, for example, each user's name, age, address, gender, skills, special abilities, etc. Access information to the user databases 2A1, 2B1, 2N1 of each server 2A, 2B, to 2N may also be stored as user attribute information in the user information database 31.
[0025] User self-assessment information includes the user's own self-assessment of characteristics such as skills and talents in each user's user attribute information. For example, the user's self-assessment may include degree information such as "not at all," "somewhat," or "completely."
[0026] User behavior history information includes, for example, information on which items each user viewed and which item icons they clicked on in the services provided by the information processing device 1. User feedback information, on the other hand, is information about the actions taken by other users on the pages that each user has made public. Here, items include, for example, heart emojis, smiley face emojis, and medal emojis. Furthermore, items such as the user's account image and images posted on social media may also be used. Since the account image set on social media is useful information, it is also expected that this account image can be used to infer the user's personality and inner self.
[0027] User feedback information includes, for example, reactions and replies from other users to certain user attribute information. Reactions include adding emojis such as smiley faces or hearts to posts, or clicking the "Like" button. Replies include direct messages such as "I have high expectations for your endless potential!", "I'm rooting for you!", "Your skills are amazing!", and "Your work has been a great help."
[0028] The word vector dictionary database 32 stores word vector dictionaries that will be used in the data preprocessing described later. A word vector dictionary (also called a "word semantic vector dictionary") is a dictionary that uses machine learning or the like to represent words as vectors and compress their dimensions in order to capture the meaning of words. The word vector dictionary allows us to measure the similarity between words as described later, and this similarity can be used to calculate peer evaluation scores, mainly for replies.
[0029] The weight information database 33 stores weight information for each characteristic item in the user attribute information, such as skills and special abilities, including weight information for the user's own self-assessment and weight information for other users' assessments of the same user. Specific examples of weight information will be described later.
[0030] The data acquisition unit 10 is configured to acquire information on each user's characteristic items from the user information database 31 of the information storage unit 30, which contains information about the user and information about the actions of other users stored in the information storage unit 30. In this embodiment, the data acquisition unit 10 acquires information on the characteristic items of the target user from the user information database 31 of the information storage unit 30 in response to operations performed by the user or other users on a mobile terminal or computer (not shown). Here, characteristic items refer to each user's skills, talents, level of interest, interests, strengths, weaknesses, etc., stored in the user information database 31. Each characteristic item is an item that can be evaluated by each user and other users. Thus, multiple characteristic items are set in the user information database 31 of the information storage unit 30 and stored as needed. Note that, as will be described later, characteristic items are also semantic categories assigned by the user themselves.
[0031] The data acquisition unit 10 is configured to perform preprocessing on the information (data) for the acquired characteristic items. As preprocessing, the data acquisition unit 10 formats the acquired text data. For example, the data acquisition unit 10 unifies the full-width / half-width characters of the text data, such as Twitter, Twitter, Twitter (registered trademark), and unifies the English / Japanese (katakana) notation. The data acquisition unit 10 also aggregates synonyms and superordinate / subordinate concepts for each acquired word. For example, for words such as citrus fruits, mandarin oranges, and oranges, the data acquisition unit 10 can aggregate the superordinate concept (citrus fruits) and the subordinate concepts (mandarin oranges, oranges). In such aggregation, a thesaurus constructed by an expert may be used to classify words based on superordinate / subordinate relationships, part / whole relationships, synonym relationships, similar relationships, etc. Building a thesaurus is costly and does not keep up with the latest vocabulary. Therefore, instead of using a thesaurus, we can measure the similarity between words using word embeddings (the positioning of words in a real vector space), and if the similarity is above a certain threshold, we can classify those words as synonyms.
[0032] The user characteristic score calculation unit 20 is configured to calculate a user evaluation score for a characteristic item based on information about each user and information about the behavior of other users regarding that characteristic item. In this embodiment, the user characteristic score calculation unit 20 calculates a user evaluation score for an arbitrary characteristic item, i.e., a tag such as skills or special abilities, using user self-evaluation information, user behavior history information, user feedback information, etc., stored in the user information database 31 of the information storage unit 30. As shown in Figure 1, the user characteristic score calculation unit 20 includes a self-evaluation score calculation unit 21, a peer evaluation score calculation unit 22, and a final evaluation score calculation unit 23.
[0033] The self-evaluation score calculation unit 21 is configured to calculate a self-evaluation score for each user based on information about each user regarding characteristic items. The self-evaluation score calculation unit 21 uses the degree of each user's own assessment of each semantic category (characteristic item), acquired by the data acquisition unit 10, directly as the self-evaluation score. For example, each user can set 1 for the highest rating and 0 for the lowest rating for each semantic category. The self-evaluation score calculation unit 21 may also be configured to scale the evaluation values according to the number of evaluation levels for the semantic category to be evaluated. For example, if the number of evaluation levels is 5, the evaluation values will be 0, 0.25, 0.5, 0.75, and 1.
[0034] The peer evaluation score calculation unit 22 is configured to calculate peer evaluation scores from other users based on information about other users' behavior regarding characteristic items. For example, the peer evaluation score calculation unit 22 can use each user's user behavior history information and user feedback information from other users to that user to link them with the semantic categories that the user has self-evaluated. In this case, the peer evaluation score calculation unit 22 calculates peer evaluation scores from other users by using the word vector dictionary stored in the word vector dictionary database 32 of the information storage unit 30 to link it with the semantic categories.
[0035] Here, we will explain an example of the operation of the peer evaluation score calculation unit 22. Figure 2 is an example of estimating the relevance of a reply considering semantic categories according to this embodiment. As shown in Figure 2, associating other users' replies with semantic categories can be done by labeling them as a document classification problem with the reply as text input. In this example, a classification model is used to classify the reply "You're working so hard! I learned a lot!" into semantic categories. In this example, for example, if the relevance of interest is 0.5 and the relevance of skills is 0.4, then the two semantic categories already equal 0.9. Therefore, this reply is estimated to be related to interest and skills. The peer evaluation score calculation unit 22 may use, for example, Naive Bayes using Bag of Words or Deep Neural Network using word embedding representations as machine learning to estimate the relevance of replies. Furthermore, since the meaning obtained by machine learning may differ depending on the context, the peer evaluation score calculation unit 22 may also embed not only the text of the reply itself, but also the context before and after the reply, or the entire context, into its input.
[0036] The final evaluation score calculation unit 23 is configured to set weight coefficients for the self-evaluation score and the peer evaluation score according to each characteristic item. The weight coefficients include a weight coefficient w1 for the self-evaluation score and a weight coefficient w2 for the peer evaluation score. One of two methods, rule-based or machine learning-based, can be used to determine these weight coefficients.
[0037] First, let's explain the rule-based weighting determination method. In the rule-based weighting determination method, the weight coefficients w1 and w2 are determined, for example, by the user (administrator) of the information processing device 1. In this case, the administrator of the information processing device 1 defines the characteristic items for which self-evaluation scores should be emphasized and the characteristic items for which peer evaluation scores should be emphasized, and then sets the weight coefficients w1 and w2 for each characteristic item based on these definitions.
[0038] Here, we will explain how to calculate the final user evaluation score (hereinafter also referred to as the "final user evaluation score") using the rule-based weighting determination method by the final evaluation score calculation unit 23. Figure 3 is an example of scoring for user attributes according to this embodiment. Here, we show the case where the user evaluation score is calculated for the user attribute "programming". For the user attribute "programming", since "interest" is determined by the user themselves, the self-evaluation score should be given more weight. Therefore, the final evaluation score calculation unit 23 sets the weight coefficient w1 for the self-evaluation score and the weight coefficient w2 for the other-evaluation score to 1 and 0.5, respectively. Also, for the user attribute "programming", since "skill" should be determined by other users, it is considered better to give more weight to the other-evaluation score. Therefore, the final evaluation score calculation unit 23 sets the weight coefficient w1 for the self-evaluation score and the weight coefficient w2 for the other-evaluation score to 0.5 and 1, respectively. In this way, the administrator of the information processing device 1 sets the weight coefficients w1 and w2 for each score for each characteristic item.
[0039] Next, the machine learning-based weight determination method will be explained. In the machine learning-based weight determination method, the weight coefficients w1 and w2 are determined, for example, based on the learned data. As initial values for the weight coefficients w1 and w2, for example, values set by the administrator of the information processing device 1 may be used, as in the rule-based weight determination method, or randomly set values may be used. The final evaluation score calculation unit 23 then uses reinforcement learning to train the model so that future user reactions (clicks by the user, reactions and replies by other users, etc.) are similar to the self-evaluation score and peer evaluation score calculated based on the currently set weight coefficients w1 and w2, and tunes (adjusts) the weight coefficients w1 and w2 at each learning timing.
[0040] For example, regarding a user's characteristic item "programming," if another user replies with "You're really good at programming," the final evaluation score calculation unit 23 might use machine learning to set the weight coefficient w2 to 0.9. Then, if yet another user replies with "This person isn't that good at programming," the final evaluation score calculation unit 23 will perform machine learning again and tune the weight coefficient w2 from 0.9 to 0.6.
[0041] Furthermore, the final evaluation score calculation unit 23 is configured to calculate a final user evaluation score for the characteristics of the user being evaluated, based on the weight coefficients, self-evaluation score, and peer evaluation score set as described above. Specifically, the final evaluation score calculation unit 23 calculates the final user evaluation score for a given characteristic as follows. (Final user evaluation score) = w1 × (self-evaluation score) + w2 × (other-evaluation score)
[0042] Here, the peer evaluation score calculated by the peer evaluation score calculation unit 22 exists for each other user who has evaluated the characteristics of a particular user. Therefore, if the peer evaluation score calculation unit 22 has calculated peer evaluation scores from multiple other users, the final evaluation score calculation unit 23 applies, for example, the average of the peer evaluation scores from multiple other users to the above formula as the overall peer evaluation score.
[0043] Furthermore, the peer evaluation score may be calculated using a method other than the average of the peer evaluation scores of multiple other users. For example, a score related to the evaluation (hereinafter referred to as the "evaluator") may be assigned to other users who have evaluated a certain user's characteristics (hereinafter also referred to as the "evaluator"), and the overall peer evaluation score may be calculated based on this score. Specifically, the evaluation-related score may be calculated by using the number of evaluations each evaluator has made, with the maximum number of evaluations by all evaluators as the denominator and the number of evaluations made by each evaluator as the numerator. This takes into account the degree of influence each evaluator has on the information processing device 1 system.
[0044] Furthermore, an evaluation-related score may be calculated by adding points to evaluators who share the same department, background, or expertise, based on the degree of relevance between each evaluator and the person being evaluated. Evaluators who share the same department are considered to have a higher accuracy in evaluating the person being evaluated, and the accuracy of the evaluation can be further improved by considering factors such as the length of time they have been in the same department. For evaluators with the same expertise, the similarity between the expertise fields of the person being evaluated and the evaluator may be measured using the word vector dictionary stored in the word vector dictionary database 32, and this similarity may be directly added to the evaluator's evaluation score, or if the similarity is above a threshold, points may be added to the evaluator's evaluation score.
[0045] Furthermore, each evaluator's evaluation-related score may be adjusted based on the error between the final user evaluation score, which is calculated using rule-based weight coefficients w1 and w2, and the final user evaluation score, which is calculated using machine learning-based weight coefficients w1 and w2, and each evaluator's score. For example, an evaluator who gives a score that is significantly different from the final user evaluation score is considered a malicious evaluator, and by lowering the score of such a malicious evaluator, the accuracy of the evaluation of the characteristics of the person being evaluated can be improved.
[0046] Furthermore, the overall peer evaluation score by other users may be calculated without using the weight coefficient w2 for peer evaluation scores set by the administrator of the information processing device 1 based on rules, or the weight coefficient w2 for peer evaluation scores adjusted based on machine learning. For example, the average of the peer evaluation scores of all evaluators may be calculated, and the standard deviation may be calculated based on the error between the average peer evaluation score and the peer evaluation score of each evaluator to determine how much they are dispersed, and each evaluator's peer evaluation score may be weighted based on this standard deviation.
[0047] In the machine learning-based weight determination method, when the final evaluation score calculation unit 23 calculates the weight coefficients w1 and w2, it simply overwrites and saves the calculated weight coefficients w1 and w2 in the weight information database 33 and uses them when calculating the final user evaluation score next time.
[0048] <Operation of the Information Processing Device> Next, an example of the operation of the information processing device 1 according to this embodiment will be described. Figure 4 is a flowchart showing an example of a scoring process performed by the information processing device 1 shown in Figure 1. This scoring process may be performed, for example, in response to an evaluation of a certain user's characteristics by another user, or in response to an operation on a mobile terminal or computer (not shown) by another user who wants to check the characteristics of that user.
[0049] When the scoring process is started, the data acquisition unit 10 acquires information for each user's characteristic items from the user information database 31 of the information storage unit 30, which includes information about the user and information about the actions of other users stored in the information storage unit 30 (step S1).
[0050] Next, the data acquisition unit 10 uses the word vector dictionary stored in the word vector dictionary database 32 to preprocess the information (data) for the acquired characteristic items, and aggregates synonyms and higher / lower concepts for each word (step S2).
[0051] Next, the self-evaluation score calculation unit 21 of the user characteristic score calculation unit 20 scores the user's self-evaluation of the characteristic items of the user being evaluated (step S3). Specifically, the self-evaluation score calculation unit 21 calculates the user's self-evaluation score based on the information about each user for the characteristic items stored in the user information database 31 of the information storage unit 30. In this case, the user characteristic score calculation unit 20 may, if necessary, use the word vector dictionary stored in the word vector dictionary database 32.
[0052] Next, the peer evaluation score calculation unit 22 of the user characteristic score calculation unit 20 scores the peer evaluations by other users for the characteristic items of the user being evaluated (step S4). Specifically, the peer evaluation score calculation unit 22 uses the user behavior history information of the user being evaluated, the user feedback information from other users to the said user, and, if necessary, the word vector dictionary database 32 to calculate the peer evaluation score by other users for the characteristic items of the user being evaluated.
[0053] Next, the final evaluation score calculation unit 23 of the user characteristic score calculation unit 20 sets weight coefficients w1 and w2 for the self-evaluation score and the peer evaluation score, respectively, according to each characteristic item. Then, the final evaluation score calculation unit 23 scores the final user evaluation based on the set weight coefficients w1 and w2, the self-evaluation score calculated in step S3, and the peer evaluation score calculated in step S4 (step S5). Specifically, the final evaluation score calculation unit 23 calculates the final user evaluation score by multiplying the self-evaluation score calculated in step S3 by weight coefficient w1, multiplying the peer evaluation score calculated in step S4 by weight coefficient w2, and adding them together.
[0054] As described above, the information processing device 1 according to this embodiment is configured to include an information storage unit 30 that stores information about a user and information about the actions of other users regarding the information about the user; a data acquisition unit 10 that acquires information about characteristic items from the information storage unit 30 among the information about the user and the information about the actions of other users regarding the characteristic items; and a user characteristic score calculation unit 20 that calculates a user evaluation score for characteristic items based on the information about the user and the information about the actions of other users regarding the characteristic items. Here, the user characteristic score calculation unit 20 is configured to include a self-evaluation score calculation unit 21 that calculates a self-evaluation score by the user based on the information about the user regarding the characteristic items; an external evaluation score calculation unit 22 that calculates an external evaluation score by other users based on the information about the actions of other users regarding the characteristic items; and a final evaluation score calculation unit 23 that sets weight coefficients for the self-evaluation score and the external evaluation score according to the characteristic items, and calculates a user evaluation score for characteristic items based on the set weight coefficients, the self-evaluation score, and the external evaluation score. By configuring the information processing device 1 in this way, it becomes possible to appropriately handle other users' evaluations of a user's characteristic items, and to correct characteristic items (evaluation items) that have been overestimated or underestimated by the user. This improves the accuracy of the evaluation of the user's characteristic items. In other words, it becomes possible to provide a more accurate user evaluation score for the user's characteristic items in the information processing device 1.
[0055] Furthermore, the information processing device 1 according to this embodiment is configured to calculate a final user evaluation score for the user's characteristic items by adjusting the weighting of subjective and objective evaluations according to the user's characteristic items. This allows the user's characteristic items to be evaluated while considering the variability and reliability of evaluations among evaluators, such as the user themselves and other users.
[0056] Thus, with the information processing device 1 according to this embodiment, the user's characteristic items (user evaluation score) are accurately evaluated. By utilizing this user evaluation score, it is possible to provide fair recommendations to third parties (users of the information processing device 1) who are looking for users with excellent characteristics. In other words, it is possible to suppress users who overestimate themselves and to identify (discover) users who underestimate themselves.
[0057] Furthermore, by comparing the user evaluation score for a particular user's characteristic item, calculated by the information processing device 1, with the peer evaluation scores of other users for the same characteristic item, it is possible to evaluate evaluators within the service based on how similar these scores are. This makes it possible to identify competent evaluators for each characteristic item.
[0058] Furthermore, in another aspect of this disclosure, an information processing method is provided. This information processing method acquires information on characteristic items from the user information database 31 of the information storage unit 30, among information on the user and information on the behavior of other users; calculates a self-evaluation score by the user based on the information on the user for the characteristic items; calculates an evaluation score by other users based on the information on the behavior of other users for the characteristic items; sets weighting coefficients for the self-evaluation score and the evaluation score by other users according to the characteristic items; and calculates a user evaluation score for the characteristic items based on the set weighting coefficients, the self-evaluation score, and the evaluation score by other users. By configuring the information processing method in this way, the same effects as the information processing device 1 described above can be achieved.
[0059] Some or all of the processing in the information processing device 1 described above can be implemented as a computer program (information processing program). Such a program can be stored using various types of non-temporary computer-readable media and supplied to a computer. Non-temporary computer-readable media include various types of tangible recording media. Examples of non-temporary computer-readable media include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). Programs may also be supplied to a computer using various types of temporary computer-readable media. Examples of temporary computer-readable media include electrical signals, optical signals, and electromagnetic waves. Temporary computer-readable media can be supplied to a computer via wired communication channels such as electric wires and optical fibers, or via wireless communication channels.
[0060] Although the present invention has been described with reference to embodiments, the present invention is not limited to the above embodiments and can be modified as appropriate without departing from the spirit of the invention. [Industrial applicability]
[0061] The information processing device described herein can be applied to improve the accuracy of evaluating a user's characteristic items by optimizing self-assessment and peer-assessment for that user's characteristic items. [Explanation of Symbols]
[0062] 1. Information Processing Device 10 Data acquisition unit 20. User Characteristics Score Calculation Unit 21 Self-evaluation score calculation unit 22. Compulsory Evaluation Score Calculation Unit 23. Final Evaluation Score Calculation Unit 30 Information storage section 31 User Information Database 32-word vector dictionary database 33. Weight Information Database 2A, 2B~2N Servers 2A1, 2B1~2N1 User Database 3 Network
Claims
1. An information storage unit that stores information about a user and information about the actions of other users regarding the information about the user, A data acquisition unit that acquires information relating to characteristic items from the information storage unit, among the information relating to the user and the information relating to the actions of other users. A user characteristic score calculation unit calculates a user evaluation score for the characteristic item based on information about the user and information about the behavior of other users for the characteristic item. Equipped with, The user characteristics score calculation unit, A self-evaluation score calculation unit calculates a self-evaluation score by the user based on the user's information regarding the characteristic items, A peer evaluation score calculation unit calculates a peer evaluation score by the other user based on information regarding the behavior of the other user with respect to the characteristic items, A final evaluation score calculation unit sets weighting coefficients for the self-evaluation score and the peer evaluation score according to the characteristic items, and calculates a user evaluation score for the characteristic items based on the set weighting coefficients, the self-evaluation score, and the peer evaluation score. Includes, The self-assessment score calculation unit calculates the self-assessment score based on the user's own self-assessment information included in the user information, The information regarding the actions of other users refers to replies made by those other users on the user pages of social networking services that the user has made public. The aforementioned peer evaluation score calculation unit calculates the peer evaluation score based on the replies. Information processing device.
2. The information storage unit stores multiple characteristic items as characteristic items. The final evaluation score calculation unit sets the weight coefficients based on rules or machine learning, respectively, according to the type of characteristic item. The information processing apparatus according to claim 1.
3. The user characteristic score calculation unit performs preprocessing on the text data acquired by the data acquisition unit using a word vector dictionary. The information processing apparatus according to claim 1 or 2.
4. A computer, Information regarding the user and information regarding the behavior of other users, specifically information related to characteristic items, is obtained from the information storage unit. Based on the information about the user regarding the aforementioned characteristic items, a self-assessment score by the user is calculated. Based on the information regarding the behavior of other users with respect to the aforementioned characteristic items, the peer evaluation score by the other users is calculated. Depending on the characteristic items, weighting coefficients for the self-assessment score and the peer assessment score are set accordingly. Based on the weight coefficients set, the self-assessment score, and the peer assessment score, a user evaluation score for the characteristic item is calculated. Includes, Calculating the self-assessment score involves calculating the self-assessment score based on the user's own self-assessment information included in the user's information. The information regarding the actions of other users refers to replies made by those other users on the user pages of social networking services that the user has made public. The calculation of the aforementioned peer evaluation score is based on the replies. Information processing methods.
5. The process involves obtaining information on characteristic items from the information storage unit, among information on the user and information on the behavior of other users. A process for calculating a self-assessment score by the user based on the user's information regarding the aforementioned characteristic items, A process for calculating a peer evaluation score by other users based on information regarding the behavior of other users with respect to the aforementioned characteristic items, A process to set weighting coefficients for the self-evaluation score and the peer evaluation score according to the aforementioned characteristic items, A process to calculate a user evaluation score for the characteristic item based on the weight coefficient set, the self-evaluation score, and the peer evaluation score, The information processing device is made to execute this, The process for calculating the self-assessment score involves calculating the self-assessment score based on the user's own self-assessment information included in the user's information, The information regarding the actions of other users refers to replies made by those other users on the user pages of social networking services that the user has made public. The process for calculating the peer evaluation score involves calculating the peer evaluation score based on the reply. Information processing program.