Personnel Evaluation System and Personnel Evaluation Method

Abstract

To provide a person evaluation system for performing person evaluation to identify a person who has characteristics of interest to an investigator and who is commonly associated with a plurality of various persons.SOLUTION: The person evaluation system maintains attribute information of each of a plurality of persons and information indicating whether each of the plurality of persons is a characteristic person having predetermined characteristics, calculates the similarity among the plurality of persons based on each attribute information of the plurality of persons, generates a cluster in which similar persons are connected among the plurality of persons based on the calculated similarity, and evaluates a specific person based on a connection relationship of the specific person included in the cluster and the characteristic person.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a person evaluation system and a person evaluation method.

Background Art

[0002] Structuring the relationships of multiple persons and evaluating the persons are required in various scenarios. For example, in registration applications for services provided by public institutions or private institutions (for example, account opening at a financial institution or real estate contract), applicants are reviewed. In this review, in addition to investigating whether the applicant meets the registration application conditions, peripheral persons related to the applicant may be investigated for risk assessment. For example, if the applicant has an acquaintance relationship with a person of concern who has been confirmed to have committed an illegal act in the past, it is considered possible that the applicant may be urged to commit an illegal act by the known person of concern, and a judgment such as conducting a focused review is made.

[0003] In addition, in a marketing survey, it may be necessary to structure and visualize relationships between persons such as interests. Based on the results, strategic planning such as identifying persons with whom partnerships should be formed for business startup is carried out.

[0004] It is possible to obtain various information from such structuring of the relationships of persons. In addition, it may be important to identify persons who are connected in common with multiple persons having characteristics that the investigator is interested in.

[0005] In the above-mentioned example of review, malicious intermediary persons and merchants who contact various persons to mediate illegal acts and obtain benefits are persons with characteristics worthy of attention. In the review, it is useful to identify whether the applicant is such a person or a person related to such a person. The peripheral persons of such malicious intermediary persons often include persons who have already been marked as persons to be noted, and have characteristics of being connected in common with various known persons of concern.

[0006] In addition, in the above example of the marketing survey, since consultants who have formed partnerships in common with high-performing businesses are likely to be excellent, identifying such consultants is useful for the success of the business. These consultants also enter into contracts individually with each business operator, and there is no clear connection between the business operators, but they have a connection in common among the multiple business operators concerned.

[0007] Here, as the background art of this technical field, there is the following prior art. Patent Document 1 (Japanese Patent Application Laid-Open No. 2012-150680) states that "in the human relationship map management server 10, in the response information storage unit 12, information on questionnaire responses from the subjects to be the creation targets of the human relationship map is stored. In the human relationship map creation support unit 100, the calculation unit 130 calculates the strength of the relationship between the subjects from the information on the questionnaire responses from the subjects. The subject identification unit 150 analyzes the calculated strength of the relationship between the subjects, extracts the subjects having a great influence on the form of the human relationship map from among the subjects who have not answered the questionnaire, and sets them as the subjects to urge for questionnaire responses. The subject output unit 170 outputs the subjects to urge for questionnaire responses." (See the abstract).

Prior Art Documents

Patent Documents

[0008]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0009] The technology described in Patent Document 1 identifies a person who has contact with various people. However, people such as malicious intermediaries and excellent consultants have a common connection with people who have a certain common characteristic, and people with that characteristic do not have an obvious connection with each other. Such a perspective is not described in Patent Document 1, and it is not possible to distinguish and extract people such as malicious intermediaries and excellent consultants from, for example, people who have a wide range of friendships within an organization.

[0010] Therefore, one aspect of the present invention performs a person evaluation for identifying a person who has a characteristic that an investigator is interested in and is commonly connected with a plurality of diverse people.

Means for Solving the Problems

[0011] A typical example of the invention disclosed in the present application is as follows. That is, a person evaluation system includes a processor and a memory. The memory holds attribute information of each of a plurality of people and information indicating whether each of the plurality of people is a characteristic person having a predetermined characteristic. The processor calculates the similarity between the plurality of people based on the attribute information of each of the plurality of people, generates a cluster in which similar people among the plurality of people are connected based on the calculated similarity, and evaluates the specific person based on the connection relationship between the specific person included in the cluster and the characteristic person.

Effects of the Invention

[0012] According to one aspect of the present invention, it is possible to perform a person evaluation for identifying a person who has a characteristic that an investigator is interested in and is commonly connected with a plurality of diverse people.

[0013] Problems, configurations, and effects other than those described above will be clarified by the description of the following embodiments.

Brief Description of the Drawings

[0014]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Embodiments for Carrying Out the Invention

[0015] Hereinafter, embodiments will be described with reference to the drawings. In the embodiments described below, the same or similar configurations or functions are denoted by the same reference numerals, and duplicate descriptions are omitted.

Embodiment

[0016] In Example 1, an example of estimating the relationship between persons based on information of a plurality of persons and extracting a person having a relationship with a diverse person having any feature from the estimated relationship will be described. This example can be used for risk assessment of persons in various review operations such as real estate and finance. Here, a person having a relationship with a diverse person having flag information of a known person of concern is extracted, and the extracted person and the surrounding persons are extracted and collated with the person to be reviewed, assuming a scenario of evaluating whether the person is a malicious intermediary or a person related thereto.

[0017] FIG. 1 is a diagram showing a configuration example of a person evaluation system in Example 1. The person evaluation system of Example 1 includes a functional unit that executes processing, and information (data) generated, updated, or used by the functional unit. The functional unit includes a similarity evaluation unit 105, a relationship network (relationship NW) construction unit 106, and a relationship network (relationship NW) analysis unit 107. The information includes person-related information 101, a similarity evaluation result 102, relationship structure information 103, and a relationship evaluation result 104.

[0018] The person evaluation system of this example is configured by a computer having a processor (CPU), a memory, an auxiliary storage device, and a communication interface. The person evaluation system may have an input interface and an output interface.

[0019] The processor is an arithmetic unit that executes a program stored in the memory. By the processor executing various programs, each functional unit provided by the person evaluation system is realized. Note that a part of the processing performed by the processor executing the program may be executed by another arithmetic unit (for example, hardware such as an ASIC or an FPGA).

[0020] The memory includes a ROM which is a non-volatile memory element and a RAM which is a volatile memory element. The ROM stores unchangeable programs (such as BIOS). The RAM is a high-speed and volatile memory element like DRAM (Dynamic Random Access Memory), and temporarily stores programs executed by the processor and data used during program execution.

[0021] The auxiliary storage device is a large-capacity and non-volatile storage device such as a magnetic storage device (HDD) or a flash memory (SSD), and stores the aforementioned information. Also, the auxiliary storage device stores programs executed by the processor. That is, the program is read from the auxiliary storage device, loaded into the memory, and executed by the processor to realize each function of the data processing system.

[0022] The communication interface is a network interface device that controls communication with other devices according to a predetermined protocol.

[0023] The input interface is an interface that receives input from the user, and the output interface is an interface that outputs the execution result of the program in a form visible to the user. Note that a user terminal connected to the person evaluation system via a network may provide the input interface and the output interface. In this case, the person evaluation system may have the function of a web server, and the user terminal may access the person evaluation system according to a predetermined protocol (such as http).

[0024] The program executed by the processor is provided to the person evaluation system via a removable medium (such as a CD-ROM or a flash memory) or a network, and is stored in a non-volatile auxiliary storage device which is a non-transitory storage medium. Therefore, it is preferable that the person evaluation system has an interface for reading data from the removable medium.

[0025] The person evaluation system is a computer system configured physically on one computer or on a plurality of computers configured logically or physically, and may operate on a virtual computer built on a plurality of physical computer resources. For example, each functional unit may operate on a separate physical or logical computer, or a plurality of them may be combined and operate on one physical or logical computer.

[0026] Figure 2 is a diagram showing a configuration example of the person-related information 101. Note that in Figures 2 to 5 and Figure 12, the configuration examples of the information in table format are shown, but other formats may also be used.

[0027] The person-related information 101 is constructed, for example, based on the data submitted in the review application. The submitted content depends on the type of review. In Example 1, it includes the basic information (such as address) of the person, the resume (such as the school of origin), and the attribute information such as the affiliated organization.

[0028] In addition to the information submitted by the person under review, the review authority may add its own information. In particular, in this embodiment, the information of known persons of concern is used, but generally, the review authority may provide information. For example, if a person reviewed in the past has committed an improper act, it is accumulated as the information of the person of concern, and a flag (flag for attention) is attached to the person included in the information of the person of concern.

[0029] Note that in this embodiment, an example where the person-related information 101 includes a flag for attention (persons of concern are managed) is described. The flag for attention is an example of a flag indicating some characteristics of the person under review. Instead of or in addition to the flag for attention, other flags indicating characteristics may be used (that is, persons with certain characteristics may be managed).

[0030] Figure 3 is a diagram showing a configuration example of the similarity evaluation result 102. The similarity evaluation result 102 stores the result processed by the similarity evaluation unit 105. The similarity is composed of one or more indicators. In this embodiment, the similarity evaluation unit 105 evaluates the similarity between persons from multiple perspectives based on each of the multiple attribute informations, and finally calculates the integrated similarity obtained by synthesizing them.

[0031] FIG. 4 is a diagram showing a configuration example of the relationship structure information 103. The relationship structure information 103 records data of a network structure. In this embodiment, it is in an explicit text format from the viewpoint of easy understanding, but it may also be a binary file or the like. The relationship structure information 103 stores, as information necessary to represent the network structure, a list of nodes (person IDs), a list of links (nodes connected by the link and the integrated similarity between the nodes connected by the link), and attribute information (attribute information used for similarity calculation, a flag to be noted, etc.) indicated by the person-related information 101.

[0032] FIG. 5 is a diagram showing a configuration example of the relationship evaluation result 104. The result processed by the relationship network analysis unit 107 is stored. In this embodiment, the final evaluation value indicating the risk of being a person to be noted and various evaluation indicators used for the determination of the final evaluation value are stored. Note that, instead of or in addition to the final evaluation value, ranking information indicating that the higher the rank, the higher the possibility of being a person to be noted may be stored.

[0033] The above is the information stored in the person evaluation system of this embodiment. Next, along with the main processing flowchart shown in FIG. 6 (a flowchart showing an example of the processing procedure of person evaluation), the processing content of each functional unit and the relationship with each input / output information will be described.

[0034] The person evaluation system receives person-related information (processing 601), and the similarity evaluation unit 105 evaluates the similarity based on the received person-related information (processing 602). An example of the detailed processing procedure of similarity evaluation will be described with reference to FIG. 7.

[0035] In the similarity evaluation process, the similarity evaluation unit 105 extracts pairs of persons for which similarity evaluation is to be performed (process 701). Since the maximum number of pairs of persons to be extracted is the total number of combinations of all two persons registered in the person-related information 101, depending on the number of persons registered in the person-related information 101 and the allowable processing time, it may be necessary to select some combinations. Also, without reducing the amount of calculation, the calculation results may be accumulated at divided calculation timings. For example, when data is registered step by step, by accumulating the similarity evaluation results of persons who have been calculated in the past, the number of pairs that need to be calculated at the timing when a new registrant appears can be reduced.

[0036] Next, the similarity evaluation unit 105 calculates the similarity (process 702). In process 702, an example of evaluating the similarity with various attribute information will be described. Some of the attribute information is described with non-continuous values such as category values and flags. For example, for attribute i of a certain person A and another person B, the similarity S based on their attributes can be defined by Expression (1). Note that the similarity S may be evaluated by other methods. Since the attention-required flag is not information indicating the attributes of a person, the similarity of the attention-required flag is not calculated.

[0037]

Number

[0038] Also, when the attribute information is a continuous value, the similarity S can be defined, for example, by Expression (2). In Expression (2), N is a constant and can be determined, for example, by the range (maximum to minimum) that attribute i can take.

[0039]

Number

[0040] Next, the similarity evaluation unit 105 evaluates the similarity for each individual attribute of one pair of persons as in process 702, and finally calculates an integrated similarity for one pair of persons as a total value (process 703). The similarity evaluation unit 105 calculates the integrated similarity, for example, by using a weighted average shown below. Each similarity S is designed to be in the range of 0 to 1, and by making the sum of each coefficient k equal to 1, the integrated similarity also becomes a numerical value in the range of 0 to 1. Note that the integrated similarity may be calculated by other methods.

[0041]

Number

[0042] The above is the detailed procedure of process 602. In this embodiment, although the process of individually evaluating and integrating the similarities for a plurality of pieces of attribute information (processes 702 and 703) is exemplified, each combination of attributes may be regarded as a multi-dimensional vector, and the similarity of the multi-dimensional vector may be calculated using cosine similarity or the like to calculate the integrated similarity.

[0043] Next, the relationship network construction unit 106 generates a network structure based on the similarity information obtained in process 602 (process 603). For example, the relationship network construction unit 106 connects pairs of persons with an integrated similarity equal to or higher than a predetermined threshold value by links, and stores the person ID (node), attribute information, attention flag, and integrated similarity between the connected persons for each cluster connected by the links in the relationship structure information 103.

[0044] Note that when persons are connected by links, one or more graphs with person IDs as nodes and links as edges are generated. At this time, all the generated graphs are collectively called a network, and each independent graph is called a cluster. That is, the nodes belonging to the same cluster are connected by a route passing through one or more links, but there is no route connecting the nodes belonging to different clusters.

[0045] Next, the relationship network analysis unit 107 analyzes the relationship network in process 604. An example of the detailed processing procedure of the relationship network analysis will be described with reference to FIG. 8.

[0046] In the relationship network analysis process, the relationship network analysis unit 107 selects one cluster from the networks obtained in process 603. Since the networks obtained up to process 603 may include a plurality of clusters, the relationship network analysis unit 107 selects one cluster from the networks. Further, the relationship network analysis unit 107 extracts the group of nodes (persons) belonging to the selected cluster (process 801).

[0047] Next, the relationship network analysis unit 107 selects one node from the group of nodes extracted in process 801. At this time, since the relationship network analysis unit 107 does not need to evaluate the persons who are already known as the persons of interest, the relationship network analysis unit 107 selects nodes other than the persons who are already known as the persons of interest (process 802). Note that the relationship network analysis unit 107 may select the persons of interest.

[0048] Next, the relationship network analysis unit 107 evaluates the node (process 803). In process 803, the relationship network analysis unit 107 evaluates the change in the configuration of the selected cluster when the selected node and the links connected to the selected node are excluded from the selected cluster.

[0049] When the cluster is divided into a plurality of clusters by excluding the node and the link, the relationship network analysis unit 107 evaluates the number of pairs of the nodes of interest whose connection relationship has been lost (that is, separated into different clusters) due to the division. The larger the number of pairs of the nodes of interest whose connection relationship has been lost, the more it is presumed that the person of the excluded node has connections with various persons of interest. Note that this index is an example of the evaluation value E1 described later.

[0050] In addition, when excluding the node, the relationship network analysis unit 107 may evaluate that the node has connections with various persons of interest by using, as an index, the number of pairs for which the increase in the number of shortest links to each person of interest node becomes equal to or more than a certain value. This index is an example of the evaluation value E1 described later. In particular, when the size of the cluster is large (the number of nodes is large), etc., since it is highly likely that the cluster will not be divided even if one node and the links connected to the node are excluded, the evaluation based on the increase in the number of shortest links is useful.

[0051] Further, the relationship network analysis unit 107 may evaluate that the node has connections with various persons of interest by the number of clusters including persons of interest among the divided clusters, by excluding the selected node and the links connected to the selected node from the selected cluster.

[0052] Also, since these indexes increase in value according to the number of persons of interest belonging to the cluster, the relationship network analysis unit 107 may normalize these indexes in order to make them easier to handle. For example, the ratio of the number of pairs whose connection relationship is lost (or the number of shortest links increases by a predetermined number or more) due to the exclusion of the node to the number of all combinations of persons of interest nodes existing in the cluster is an index normalized to the range from 0 to 1. Thereby, an index representing the contribution degree to the connection between persons of interest in the cluster to which the node belongs is defined. This index is an example of the evaluation value E2 described later.

[0053] The relationship network analysis unit 107 may perform evaluation using other indexes related to the node in addition to or instead of the above-described indexes. For example, the relationship network analysis unit 107 performs evaluation using the relationship strength between persons calculated by utilizing the integrated similarity. For persons connected via a plurality of links, the relationship network analysis unit 107 calculates the product of the integrated similarities for each of all possible paths, and determines the relationship strength between the persons by the maximum value among them.

[0054] As a result, the relationship network analysis unit 107 can calculate the relationship strength between the node selected in step S802 and each of all the nodes in the cluster. Among these, the relationship network analysis unit 107 obtains the maximum value of the relationship strength between the node selected in step S802 and the person of interest in the cluster. The maximum value is an index indicating the strength of the relationship with the person of interest closest to the node.

[0055] After calculating the above-mentioned one or more indicators, the relationship network analysis unit 107 performs a final risk assessment. For example, the relationship network analysis unit 107 determines the weighted average of the above-mentioned one or more indicators as the final evaluation value.

[0056] For example, for person A, evaluation value E1 A is the contribution degree of person A to the connection between persons of interest in the cluster to which person A belongs, and evaluation value E2 A is the maximum value of the relationship strength between person A and the person of interest. Evaluation value E1 A and evaluation value E2 A are also evaluation values based on the connection relationship between the person under review and the person of interest belonging to the same cluster as the person under review. K i is a weight coefficient, and the sum of K i is 1. As a result, the following final evaluation value is calculated within the range of 0 to 1.

[0057]

Equation

[0058] Note that each weight coefficient K i is adjusted according to the perspective emphasized in the risk assessment. The larger K1 is, the higher the evaluation of a person with a strong nature as an intermediary connecting various persons of interest can be. The larger K2 is, the higher the evaluation of a person with a close connection to a specific person of interest can be.

[0059] In the above example, the attention flag is given from outside the person evaluation system. However, the relationship network analysis unit 107 may calculate (update) the attention flag. Specifically, for example, the relationship network analysis unit 107 may update the attention flag of a person whose calculated final evaluation value is equal to or greater than a predetermined value to "1" (i.e., a value indicating that the person is a person of concern), and update the attention flag of a person whose calculated final evaluation value is less than the predetermined value to "0" (i.e., a value indicating that the person is not a person of concern).

[0060] Subsequently, the relationship network analysis unit 107 executes processes 802 and 803 for each node of the evaluation target among the node groups extracted in process 801 (process 804). The relationship network analysis unit 107 executes processes 801 to 804 for each cluster of the evaluation target (process 805).

[0061] The relationship network analysis unit 107 outputs the processing result as a relationship evaluation result to a display or a printed matter (process 806). Note that the network may be a group of several hundred people, and it may be difficult for the user to view the whole as it is visualized. Therefore, it is desirable to perform processing such as reducing the amount of information according to the application and narrowing down the display range.

[0062] FIG. 9 is an example of a display screen of the analysis result when the person to be examined is determined in the examination. On the display screen of FIG. 9, for example, a network 901, an examinee summary 902, and peripheral person information 903 are displayed.

[0063] The network 901 visualizes the person network up to the link destination specified by the user of the person evaluation system with the examinee as the center. In the network 901, the nodes indicating persons who are not persons of concern are represented in white, and the nodes indicating persons of concern are represented in black. Also, the length of the link may be the same predetermined value for all, or may be shorter as the integration similarity between the nodes connected by the link is higher.

[0064] The examinee summary 902 shows the evaluation results of the examinee (for example, the final evaluation value, E1, E2, etc. used in calculating the final evaluation value). The peripheral person information 903 shows the evaluation results of the persons extracted as the peripheral persons of the examinee. The peripheral persons of the examinee may be, for example, persons belonging to the same cluster as the examinee, persons connected via a link within a predetermined number from the examinee, or persons with a relationship strength with the examinee equal to or greater than a predetermined value.

[0065] In the peripheral person information 903, for example, a flag to be noted for the peripheral person, the relationship strength between the peripheral person and the examinee, and attribute information (for example, attribute information with a similarity equal to or greater than a predetermined value, etc.) that contributed to the improvement of the integrated similarity in the similarity evaluation are displayed.

[0066] FIG. 10 is an example of a display screen of the analysis result when checking whether there are persons with high risks from the person relationship in regular inspections or the like. Even for persons who have been evaluated in past reviews, if information related to the person relationship is updated, such as when a person to be noted is newly registered later, there is a possibility that the risk will increase compared to the past review stage when the evaluated person is re-evaluated. Therefore, the analysis process may be executed not only at the review stage but also thereafter (for example, at regular intervals or every time a new person to be noted is registered).

[0067] On the display screen of FIG. 10, for example, a person list 1001 and a network 1002 are displayed. In the person list 1001, for example, persons are listed in descending order of risk (in descending order of the final evaluation value). When the checkbox of the person in the person list 1001 is selected, the network 1002 centered on the selected person is displayed.

[0068] As described above, the person evaluation system of Example 1 constructs a network showing the relationships between persons based on the integrated similarity between persons calculated based on various attribute information of the persons. Further, the person evaluation system evaluates the person based on the contribution degree of the person to the connection between known persons of interest (characteristic persons) in the cluster to which the person belongs and the evaluation value based on the relationship strength between the person and the known person of interest, so that it is possible to estimate whether the person has a deep relationship with the known person of interest, whether there is a connection with various persons and groups, and the like.

Example

[0069] The person evaluation system of Example 2 estimates the relationship between persons based on information about a plurality of persons, and extracts persons having relationships with various persons having any known characteristics based on the estimated relationship. The person evaluation system of Example 2 is applicable, for example, to the risk determination of persons in various review operations such as real estate or finance.

[0070] The processing by the person evaluation system of Example 2 differs from the processing in Example 1 in that, in the final evaluation, it can output not only the presence or absence of the relationship between persons but also the determination result of the type of relationship.

[0071] In Example 2, relationship information such as the presence or absence of a relationship between persons and the type of relationship may be partially obtained in advance. For example, depending on the review application form, there may be columns for entering family members, cohabitants, guarantors, etc., and person information may be recorded. Also, in the information of social networking services, relationship information and its tag (family, friend, etc.) information may be managed. The person evaluation system may also perform evaluation using such known relationship information.

[0072] FIG. 11 is a diagram showing a configuration example of the person evaluation system in Example 2. The person evaluation system of Example 2 further includes, as information, known person relationship information 108 and relationship type dictionary 109 in addition to the configuration of Example 1.

[0073] 12 is a diagram showing an example of the configuration of known person relationship information 108. The person relationship information 108 stores, for each person, the person ID of a person with whom the person has a relationship and information indicating the relationship between the two people.

[0074] Fig. 13 is a diagram showing an example of the configuration of the relationship type dictionary 109. Fig. 13 shows an example in which the relationship type dictionary 109 is in an explicit text format for ease of understanding, but it may be a binary file or the like. The relationship type dictionary 109 includes, for example, a conditional expression for determining a relationship based on a condition of attribute information. For example, the person-related information 101 defines the company and department to which a person belongs, and the relationship type dictionary 109 defines an if-then rule such as determining that the person is a colleague if the company and department to which the person belongs match. Note that the conditional expression is not limited to this format.

[0075] The process executed in the second embodiment is the same as that in the first embodiment, that is, the processes in Figs. 6, 7, and 8 are executed. However, some processes are different. The following describes the differences between the processes executed in the second embodiment and those executed in the first embodiment.

[0076] The relationship network analysis unit 107 assigns relationship information between the person selected in process 802 and surrounding persons or persons of caution in the node evaluation of process 803. When known person relationship information between the person selected in process 802 and surrounding persons or persons of caution is given, the relationship network analysis unit 107 stores the information in the relationship evaluation result 104.

[0077] If no known personal relationship information is given between the selected person and surrounding persons or persons of caution, and a relationship that satisfies a conditional expression in the relationship type dictionary 109 is recognized between the selected person and surrounding persons or persons of caution, the relationship network analysis unit 107 stores the relationship information that satisfies the conditional expression in the relationship evaluation result 104.

[0078] When the relationship between the selected person and the surrounding people or persons of interest does not correspond to any of the above, the relationship network analysis unit 107 stores fixed information such as "undetermined" as the relationship information between the selected person and the surrounding people or persons of interest in the relationship evaluation result 104.

[0079] In addition, the relationship network analysis unit 107 may display the relationship information stored in the relationship evaluation result 104 as one column of the evaluation result on the display screen in FIG. 9.

[0080] As described above, the person evaluation system according to the second embodiment can visualize the relationship between persons by assigning relationship information between persons using the known person relationship information 108. In addition, the person evaluation system according to the second embodiment can infer and visualize the relationship between persons from the attribute information of the persons and the relationship type dictionary 109.

[0081] Note that the present invention is not limited to the above-described embodiments, and includes various modifications and equivalent configurations within the scope of the appended claims. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and the present invention is not necessarily limited to those having all the configurations described. Also, part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Further, the configuration of another embodiment may be added to the configuration of one embodiment. Also, for a part of the configuration of each embodiment, addition, deletion, or replacement of other configurations may be made.

[0082] In addition, each of the above-described configurations, functions, processing units, processing means, etc. may be realized in hardware by designing part or all of them, for example, by using an integrated circuit, or may be realized in software by a processor interpreting and executing a program for realizing each function.

[0083] Information such as programs, tables, and files for realizing each function can be stored in a storage device such as a memory, a hard disk, an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

[0084] In addition, control lines and information lines show those considered necessary for explanation purposes, and do not necessarily show all control lines and information lines required for implementation. In reality, it is reasonable to consider that almost all components are interconnected.

Explanation of Signs

[0085] 101 Person-related information, 102 Similarity evaluation result, 103 Relationship structure information, 104 Relationship evaluation result, 105 Similarity evaluation unit, 106 Relationship network construction unit, 107 Relationship network analysis unit, 108 Known person relationship information, 109 Relationship type dictionary

Claims

1. A person evaluation system, comprising a processor and a memory, wherein the memory holds attribute information of each of a plurality of persons and information indicating whether each of the plurality of persons is a characteristic person having a predetermined characteristic, and the processor calculates a similarity between the plurality of persons based on the attribute information of each of the plurality of persons, generates a cluster connecting similar persons among the plurality of persons based on the calculated similarity, evaluates the specific person based on a connection relationship between the specific person included in the cluster and the characteristic person, and evaluates the specific person based on a change in the configuration of the characteristic person of the cluster when the specific person is excluded from the cluster. A person evaluation system.

2. The person evaluation system according to claim 1, wherein the processor evaluates the specific person based on a change in the presence or absence of a connection relationship between the characteristic persons included in the cluster when the specific person is excluded from the cluster. A person evaluation system.

3. The person evaluation system according to claim 1, wherein the processor evaluates the specific person based on a change in the number of connection paths between the characteristic persons included in the cluster when the specific person is excluded from the cluster. A person evaluation system.

4. The person evaluation system according to claim 1, wherein the processor evaluates the specific person based on the number of clusters including the specific person among the divided clusters when the specific person is excluded from the cluster. A person evaluation system.

5. A person evaluation system, comprising a processor and a memory, wherein the memory holds attribute information of each of a plurality of persons and information indicating whether each of the plurality of persons is a characteristic person having a predetermined characteristic, and the processor calculates a similarity between the plurality of persons based on the attribute information of each of the plurality of persons, generates a cluster connecting similar persons among the plurality of persons based on the calculated similarity, evaluates the specific person based on a connection relationship between the specific person included in the cluster and the characteristic person, and generates data for outputting to a display device an evaluation value of the specific person based on the connection relationship between the specific person included in the cluster and the characteristic person and information indicating the characteristic person included in the cluster. A person evaluation system.

6. The person evaluation system according to claim 1 or 5, wherein the processor calculates the strength of the relationship between the specific person and each of the characteristic persons included in the cluster based on the calculated similarity, and evaluates the specific person based on the maximum strength of the calculated relationships. A person evaluation system.

7. The person evaluation system according to claim 1 or 5, wherein the memory holds person relationship information indicating the relationships between the plurality of persons and condition information for defining the relationships between persons based on the attribute information, and the processor infers the relationships between the persons included in the plurality of persons based on the attribute information of the plurality of persons and the condition information, and generates data for outputting the relationships indicated by the person relationship information and the inferred relationships. A person evaluation system.

8. A person evaluation method by a person evaluation system, wherein the person evaluation system includes a processor and a memory, the memory holds the attribute information of each of the plurality of persons and information indicating whether each of the plurality of persons is a characteristic person having a predetermined characteristic, and the person evaluation method includes the processor calculating the similarity between the plurality of persons based on the attribute information of each of the plurality of persons, the processor generating a cluster connecting the similar persons among the plurality of persons based on the calculated similarity, the processor evaluating the specific person based on the connection relationship between the specific person included in the cluster and the characteristic person, and the processor evaluating the specific person based on the compositional change of the characteristic person of the cluster when the specific person is excluded from the cluster. A person evaluation method.

Images