JP2021117583A - Compatibility decision program using model's interaction part, device and method, and matching program - Google Patents

Abstract

To provide a program that can determine compatibility between objects by considering characteristics of each object.SOLUTION: This program is a compatibility decision program for determining information pertaining to compatibility between objects having predetermined characteristics expressed in at least one factor score, and makes a computer function as a means of determining a compatibility score that determines a compatibility score pertaining to compatibility between one object and the other object for an interaction part in the pre-built model with the interaction part for the first explanatory variable, which is the factor score of the first target, and the second explanatory variable, which is the factor score of the second target, where the result value, which is the value pertaining to the result of the action of the first object and the second object, is used as the objective variable, based on the value obtained by applying the factor score of one object and the factor score of the other object.SELECTED DRAWING: Figure 1

Description

The present invention relates to a technique for estimating compatibility between objects based on characteristics such as personality.

There are frequent occasions when people form groups, such as English conversation classes, in-house projects, organizations related to human activities such as sports teams, and temporary groups formed by call center operators and customers, for example. .. At this time, it is also empirically known that the so-called "compatibility" between the members greatly influences the performance such as the performance and satisfaction in the group.

Therefore, it is highly expected that the performance of the group will be improved if the matching grouping is carried out in consideration of compatibility. Here, it has been pointed out as some research results that the personality of each person is cited as an important factor that determines the performance of the group.

For example, Non-Patent Document 1 discloses an experimental result that a direct correlation was found between the performance of two users and the personality of each user. Specifically, for example, the higher the extroversion and cooperation of one of the users, the more satisfied the travel satisfaction of both users is.

Similarly, Non-Patent Document 2 discloses an experimental result showing a strong correlation between the performance of two users and the personality of each user. Here, for example, it is verified that the higher the tendency of neurosis in any one user, the lower the satisfaction level in both users. Furthermore, Non-Patent Document 3 discloses that the similarity of personality between two users shows a correlation with compatibility.

D. P. H. Barelds, “Self and partner personality in intimate relationships”, European Journal of Personality, vol.19, no. 6, pp.501-518, 2005 JP Caughlin, TL Huston, and RM Houts, “How does personality matter in marriage? An examination of trait anxiety, interpersonal negativity, and marital satisfaction”, Journal of Personality and Social Psychology, vol.78, no. 2, pp.326 -336, 2000 P. Dijkstra and D. P. H. Barelds, “Do People Know What They Want: A Similar or Complementary Partner?”, Evolutionary Psychology, vol.6, no. 4, pp. 595-602, 2008 Toshihiro Kamishima. "Algorithms of Recommender Systems (1)", Journal of the Japanese Society for Artificial Intelligence, Vol. 22, No. 6, pp. 826-837, 2007

However, even with the prior art as described in Non-Patent Documents 1 to 4, it is still impossible to estimate the compatibility of two users with the personalities of both.

For example, although the techniques described in Non-Patent Documents 1 and 2 suggest a correlation between the performance of two users and the personality of each user, no mention is made of the relationship between the personalities of the two users. Not. Moreover, in fact, no direct estimation of the compatibility of these users has been performed.

In addition, the technique described in Non-Patent Document 3 certainly captures that the similarity of personality between two users shows a correlation with compatibility. However, it does not even refer to the relationships between individual personality factors, and therefore, it has not yet been possible to directly estimate the compatibility of these users with their individual personalities.

Incidentally, the technique described in Non-Patent Document 4 relates to a system that estimates and recommends content that matches the user. As described above, in the matching between a person and a product / thing, a large number of matching case data as correct answers can usually be obtained, so that the estimation is relatively easy. This point can be easily understood from the fact that, for example, on a certain video sharing site, there are a large number of users who have watched 30 or more videos in a predetermined period.

On the other hand, in matching between people, it is usually difficult to acquire a large number of matching case data per person. This can be understood from the fact that, for example, users who have been taught by 30 teachers in an English conversation class are extremely rare. As a result, the matching result matrix showing the matching status between people is related to a space that is extremely sparse compared to that of people and products. For example, technologies such as collaborative filtering commonly used in recommendation technology It becomes difficult to apply.

In this way, matching between people is difficult due to the scarcity of case data per person, but from this as well, estimating compatibility between people is not possible with conventional technology. It is understood that it is not easy at all.

Therefore, an object of the present invention is to provide a compatibility determination program, apparatus and method capable of determining compatibility between objects by considering the characteristics of each object. Another object of the present invention is to provide a matching program capable of performing matching between objects using the compatibility determined in this way.

According to the present invention, it is a compatibility determination program that functions a computer that determines compatibility information between objects having a predetermined characteristic expressed by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. The factor score of one target and the factor score of the other target for which the compatibility is determined with respect to the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the target. A compatibility determination program is provided in which a computer functions as a compatibility score determining means for determining the compatibility score relating to the compatibility between the one object and the other object based on the value obtained by applying the above.

As an embodiment of the compatibility determination program according to the present invention, the compatibility determination program is
The first target group is used by using the result value obtained for each combination of each target of the first target group to which the one target belongs and each target of the second target group to which the other target belongs. For each combination of each factor score of the subject belonging to the second target group and each factor score of the subject belonging to the second target group, a value indicating the significance of the interaction between the factor scores related to the combination is calculated. Sex analysis means and
Model construction that determines an interaction part having only a term for the factor score related to the combination judged to be significant based on the value indicating the significance, and constructs the model having the determined interaction part. It is also preferable to make the computer function further as a means.

Further, as a specific example in the compatibility determination program according to the present invention, the target is a person, the predetermined characteristic is a personality characteristic, and the result value is a person who is the first target and a second target. It is also preferable that it is a value related to a score, rank or rank given to the result of a predetermined act performed by either one of the persons alone or jointly.

According to the present invention, matching between objects belonging to each group in the first object group and the second object group, which is a set of objects having a predetermined characteristic expressed by at least one factor score, is performed. A matching program that makes the computer that decides work
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. For the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the object, each factor score of the object belonging to the first object group and the second object Based on the value obtained by applying the combination with each factor score of the subject belonging to the group, for each combination of each subject belonging to the first target group and each subject belonging to the second target group, the combination of the combination A compatibility score determining means for determining compatibility scores related to compatibility, and
A matching program that functions a computer as a matching means for determining a target belonging to a second target group that matches a target belonging to the first target group based on a compatibility score determined for each combination is provided. NS.

As one embodiment of the matching program according to the present invention, in the matching means, each target belonging to the first target group is set as a first node, and each target belonging to the second target group is set as a second node. Generates a directed bipartite graph connecting the node and the second node with an edge, and the cost of the edge is the target corresponding to the first node related to the edge and the second node related to the edge. After determining based on the compatibility score determined for the combination with the target corresponding to, the bipartite graph is solved as the minimum cost flow problem, and the second part that matches the target belonging to the first target group is matched. It is also preferable to determine the target belonging to the target group.

Here, in the above embodiment, the matching means includes an edge connecting the set source and the first node, and an edge connecting the second node and the set sink, respectively, according to the bipartite graph. It is also preferable to determine one target belonging to the second target group that matches one target belonging to the first target group by solving the minimum cost flow problem with the maximum flow rate set to 1 for the edge.

Alternatively, in the above embodiment, the matching means solves the minimum cost flow problem by setting the maximum flow rate of each edge connecting the set source and the first node to N (N is an integer of 2 or more). It is also preferable to determine N objects belonging to the second object group that match one object belonging to one object group.

According to the present invention, there is also a compatibility determination device for determining compatibility information between objects having a predetermined characteristic represented by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. The factor score of one target and the factor score of the other target for which the compatibility is determined with respect to the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the target. Provided is a compatibility determination device having a compatibility score determining means for determining the compatibility score relating to the compatibility between the one object and the other object based on the value obtained by applying the above.

According to the present invention, there is a compatibility determination method in a computer for determining compatibility information between objects having a predetermined characteristic represented by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. Steps to build a model with an interaction with the second explanatory variable, which is the factor score of interest,
The one object and the other object are based on the values obtained by applying the factor score of one object and the factor score of the other object, which are the compatibility determination targets, to the interaction part in the constructed model. A compatibility determination method is provided that includes a step of determining a compatibility score relating to compatibility with.

According to the compatibility determination program, apparatus and method of the present invention, the compatibility between objects can be determined by considering the characteristics of each object. Further, according to the matching program of the present invention, matching between objects can be performed using the compatibility determined in this way.

It is a functional block diagram which shows the functional structure in one Embodiment of the matching apparatus which includes the compatibility determination apparatus by this invention. It is a schematic diagram for demonstrating one Embodiment of the interfactor significance analysis processing in the interfactor significance analysis part which concerns on this invention. It is a schematic diagram for demonstrating one Embodiment of the model construction process and the compatibility score determination process in the model construction unit and the compatibility score determination unit which relate to this invention. It is a schematic diagram for demonstrating one Embodiment of the matching process in the matching part which concerns on this invention. It is a schematic diagram for demonstrating another embodiment of the matching process in the matching part which concerns on this invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[Compatibility determination device, matching device]
FIG. 1 is a functional block diagram showing a functional configuration in one embodiment of a matching device including a compatibility determination device according to the present invention.

The matching device 1 of the present embodiment shown in FIG. 1 includes a function as a compatibility determination device according to the present invention.
(A) A predetermined characteristic expressed by one or more factors in a person to be determined for compatibility, for example, each teacher and each student of an English conversation school by communication, for example, from an external personality information database (DB) 2. Get personality information, which is information, and then
(B) A result value that is a value related to the result of the action of the compatibility determination target persons, for example, by communication from the performance information DB3 installed outside, for example, a grade or a test result in a certain student who took a lesson from a certain teacher. Get some performance information and
(C) A regression model is formed using the information acquired in (a) above as an explanatory variable and the information acquired in (b) above as an objective variable, and the compatibility between the compatibility determination target persons is determined based on this model. In addition, matching is performed between the compatibility determination target persons.

Here, in the personality information DB2, for example, the scores of each factor in the five major factors (Big Five) of each teacher and each student, which are judged and determined by the questionnaire survey conducted for the teacher and the student, are the target IDs. It may be recorded in association with. By the way, the five major factors (Big Five) are personality information proposed and widely accepted in the FFM (Five Factor Model) proposed by Goldberg et al. It is composed of five factors: C), extroversion (E), harmony (A), and emotional instability (N).

On the other hand, in the performance information DB3, for example, for each teacher / student pair in these teachers and students, or for each class in charge of the teacher, the performance information which is the student's grades and test results is stored in the related teachers and students. It may be recorded in a form associated with the target ID. The personality information DB2 and the performance information DB3 can be stored or connected to the same server managed by the English conversation school, and managed by a common target ID, for example.

Further, in the present embodiment, the objects related to the information recorded and stored in the personality information DB2 and the performance information DB3 are classified into a first target group (for example, a teacher group) and a second target group (for example, a student group). It is assumed that there is. The matching device 1 determines the matching between the objects belonging to the respective groups (for example, between the teacher and the student) in the first object group and the second object group.

Specifically, the matching device 1
(A1) A "model" whose objective variable is a result value (for example, performance information such as grades), which is a value related to the result of an action between a first object (for example, a teacher) and a second object (for example, a student). The first explanatory variable, which is the factor score of the first target (for example, the teacher's Big Five score set), and the second explanatory variable, which is the factor score of the second target (for example, the student's Big Five score set). For this "interaction part" in a pre-built "model" with an "interaction part" for
(A2) A combination of each factor score of the subject belonging to the first target group (for example, each Big Five score of the teacher) and each factor score of the subject belonging to the second target group (for example, each Big Five score of the student). For each combination of each subject (each teacher) belonging to the first target group and each subject (each student) belonging to the second target group, based on the value obtained by applying The compatibility score determination unit 113 that determines the "compatibility score" related to
(B) Matching to determine the target (student) belonging to the second target group that matches the target (teacher) belonging to the first target group based on the "compatibility score" determined for each combination. It is characterized by having a portion 114.

Here, the "interaction part" in the "model" of the above (A1) is a model part indicating that the influence of each of the first explanatory variable and the second explanatory variable on the objective variable depends on each other's values. In this embodiment, as will be described in detail later, it is a part corresponding to the interaction term of the multiple regression equation when the "model" is expressed by the multiple regression equation.

The inventors of the present application grasp that such interaction of the "interaction part" is evaluated based on the result value (for example, performance information such as results) due to the action of both, as a suitable / effective degree of the combination of the two. Therefore, it was found that it corresponds to the so-called "compatibility", and the "compatibility score" between the subjects was determined from this "interaction part". By the way, as an easy-to-understand scene of such interaction in "compatibility", for example, it is expected that a teacher with a high extroversion will generally improve the student's grades, but such a teacher is sincere. It is possible that students with high sexuality (planning) may feel annoyed.

As described above, according to the matching device 1, it is possible to determine the "compatibility" between the objects by considering the interaction of the characteristics (for example, personality information) of each object. Further, in the present embodiment, the matching device 1 can perform matching between objects by using the "compatibility" determined in this way. Here, a specific method of this matching will be described in detail later with reference to FIGS. 4 and 5.

The matching device 1 is not limited to the "compatibility" between the target of the first target group (for example, a teacher) and the target of the second target group (for example, a student), for example, between the targets of the first target group (for example). Of course, it is also possible to determine the "compatibility" between the subjects (for example, between teachers) and the "compatibility" between the subjects (for example, between the students) of the second target group. That is, the matching device 1 can estimate the "compatibility" between the two objects from the characteristics of each of the two objects.

Further, the result value of the above (A1) is not limited to performance information such as grades and test results, and various values related to the result due to the action between the objects can be adopted as the result value. Is. For example, it is possible to adopt a value indicating the state of the athlete who has been instructed by the trainer, for example, a weight, a chest circumference, an abdominal circumference, a numerical value indicating physical ability / flexibility, and the like as a result value. Further, as this result value, a score given to the result of a predetermined action performed by either one of the first target person and the second target person alone or jointly. It is also preferable to adopt a value related to rank or rank.

Further, the "characteristic" expressed by the factor score of (A1) above is not limited to the personality. For example, it is also possible to adopt the cognitive ability (intelligence quotient IQ, etc.) and non-cognitive ability (GRIT, emotional intelligence quotient EQ, etc.) possessed by the subject. However, in this embodiment, in particular, performance information as an objective variable is adopted as a characteristic related to an explanatory variable that better explains it.

Furthermore, the targets for which compatibility determination / matching processing should be performed are not limited to teachers and students. For example, it is possible to determine compatibility and matching between a call center operator and a user (customer). In this case, it is expected that the performance of the call center will be improved by predicting the compatibility between the operator and the user in advance and assigning an appropriate operator to the user.

In fact, operators' customer service skills are not constant, and performance performance varies. Therefore, since it is impossible to assign highly skilled operators to all users, it is possible to greatly contribute to the improvement of overall performance by allocating operators in consideration of compatibility with users.

Here, although it partially overlaps with the above, as specific examples of the result value of (A1) above, the student's academic ability test result in charge of the teacher, the student's confidence in the teacher, and the teacher in the class to the student Stress level, contract rate obtained by the operator from the customer, customer satisfaction with the operator, response time required by the operator to the customer, contract rate of male and female pairs in marriage activity matching, men and / or women who participated in marriage activity matching Various numerical values and quantities that should be considered in business management, such as the degree of satisfaction of the above, the number of complaints from men and / or women who participated in the marriage hunting matching, can be listed.

Further, the compatibility score determination unit 113 of the matching device 1 (A1) and (A2) and the matching unit 114 of the matching device (B) are included in separate devices (that is, the compatibility determination device and the matching device). It is also possible to take various forms. Furthermore, it is also possible that the interfactor significance analysis unit 111 and the model construction unit 112, which are the model construction function units described in detail later, are separate devices. Here, even in such a case, the whole of these can be regarded as a compatibility determination / matching device or system that implements the compatibility determination / matching method according to the present invention.

[Device function configuration, program, compatibility determination method, matching method]
Similarly, according to the functional block diagram of FIG. 1, the matching device 1 including the function as the compatibility determination device includes the communication interface unit 101, the personality information storage unit 102, the performance information storage unit 103, and the compatibility score storage unit 104. It has a matching result storage unit 105, a keyboard (KB) 106, a display (DP) 107, and a processor memory.

Here, this processor memory stores one embodiment of a matching program including the function of the compatibility determination program according to the present invention, and also has a computer function, and by executing this matching program, Perform compatibility determination processing and matching processing. For this reason, the matching device 1 may be a server, a cloud server, or a dedicated device for matching, but is equipped with a matching program according to the present invention, for example, a personal computer (PC), a notebook computer, or a tablet computer. , Or a smartphone or the like.

Further, the processor memory includes an interfactor significance analysis unit 111 including an analysis of variance unit 111a, a model construction unit 112, a compatibility score determination unit 113 including a multiple regression analysis unit 113a, and a directed bipartite graph generation. It has a matching unit 114 including the unit 114a, a communication control unit 121, and an input / output control unit 122. Note that these functional components can be regarded as the functions of the matching program stored in the processor memory. Further, among these, the interfactor significance analysis unit 111, the model construction unit 112, and the compatibility score determination unit 113 can be regarded as the functions of the compatibility determination program included in the matching program. Further, the processing flow shown by connecting the functional components of the matching device 1 in FIG. 1 with arrows is also understood as an embodiment of the matching method including the compatibility determination method according to the present invention.

Similarly, in the functional block diagram of FIG. 1, the personality information storage unit 102 and the performance information storage unit 103 receive personality information (for example, from the personality information DB2 and the performance information DB3, respectively, via the communication interface unit 101 and the communication control unit 121, respectively. Save and manage Big Five score sets for each teacher and each student) and performance information (for example, test results for each combination of teacher and student).

Here, the personality information (Big Five score) stored in the personality information storage unit 102 is, for example, a program, an apparatus, and a method for estimating a personality characteristic from a browsing history of a user's page, for example, Patent Document: Japanese Patent Application Laid-Open No. 2019-197338. It may be estimated by the technique described in.

Further, in this case, for example, the matching device 1 includes a functional unit that acquires browsing history information of a user's web page from a web (Web) server and estimates the personality information of the user from the browsing history information. It is also preferable to generate personality information by oneself. Furthermore, the matching device 1 may be, for example, a stand-alone device that acquires personality information and performance information by a method other than communication.

Similarly, in the functional block diagram of FIG. 1, the interfactor significance analysis unit 111
(A) Each target of the first target group (for example, teacher group) to which one target (for example, a teacher) to which the compatibility is determined belongs, and the second target group (for example, student group) to which the other target (for example, a student) belongs. The result values (for example, test results) obtained for each combination with each target of) are acquired from the performance information storage unit 103, and these result values are used.
(B) Each factor score (for example, Big Five score) of the object (teacher) belonging to the first target group and each factor score of the object (student) belonging to the second target group obtained from the personality information storage unit 102. For each combination with (for example, Big Five score), a value indicating the significance of the interaction between the factor scores (Big Five score) related to the combination, and the P value in the present embodiment are calculated.

Here, in the present embodiment, the two-way analysis of variance unit 111a of the interfactor significance analysis unit 111 performs the two-way analysis of variance, which is a known method, and calculates the P value Pr for each combination. Regarding two-way analysis of variance, for example, non-patent literature: Ikuo Ikeda, "For those who use statistical tests without understanding III" Chemistry and Biology, vol.51, no. 7, pp.483- Explained in detail in 495, 2013.

The P value calculated by the two-way analysis of variance unit 111a is subsequently used to determine the form of interaction of the interaction unit (interaction term) in the performance estimation model to be constructed. That is, in reality, if an attempt is made to perform an interaction term estimation process using a model that considers all combinations of factor scores (Big Five scores) between the two (teacher and student), the number of dimensions in the process becomes enormous, and regression occurs. May become difficult. Therefore, in the present embodiment, only the combination of the factors determined to be significant is incorporated into the interaction term by using the calculated P value.

FIG. 2 is a schematic diagram for explaining an embodiment of the interfactor significance analysis process in the interfactor significance analysis unit 111.

According to FIG. 2, the interfactor significance analysis unit 111 first
(A) Obtain the Big Five scores of a plurality of teachers (target ID = teacher 01, teacher 02, teacher 03, ...) belonging to the target group 1 from the personality information storage unit 102.
(B) Obtain the Big Five scores of a plurality of students (target ID = student 01, student 02, student 03, ...) belonging to the target group 2 from the personality information storage unit 102.
(C) From the performance information storage unit 103, all the test results of the students of the target group 2 in charge of the teacher of the target group 1, that is, the test results for each combination of the teacher of the target group 1 and the students of the target group 2 are acquired. do.

Next, the interfactor significance analysis unit 111 determines the score of the teacher's Big Five factor and the student's Big Five factor for each combination of the teacher's Big Five factor belonging to the target group 1 and the student's Big Five factor belonging to the target group 2. Generate a table in which the scores of the Big Five factors of the above are recorded in association with the test results of the student in charge of the teacher. Note that FIG. 2 shows a table relating to the combination of the "extroversion" of the teacher belonging to the target group 1 and the "integrity" of the students belonging to the target group 2.

Further, the two-way analysis of variance unit 111a of the interfactor significance analysis unit 111 performs a two-way analysis of variance processing on the data in the table generated for each combination, and for each combination, the teacher's Big Five The P-value Pr in the interaction between the factor and the student's Big Five factor is calculated. Note that FIG. 2 also shows various values other than the P value Pr of the interaction obtained by the two-way analysis of variance processing.

Here, the P value Pr of the interaction is a value indicating the significance of the interaction, and specifically, under the assumption that the null hypothesis that there is no difference in the test results due to the interaction is correct. It is a value corresponding to the probability that a difference in test results will occur. Further, in the field of statistical analysis, it is customary to reject the null hypothesis if the P value is 0.05 (5%) or less (as a significance level), and therefore, in the present embodiment, the P of this interaction When the value Pr is 0.05 (5%) or less, the interaction between the teacher's Big Five factor and the student's Big Five factor is judged to be "significant". By the way, the P value Pr (0.0040287) of the interaction shown in FIG. 2 is below the significance level (0.05), and as a result, the teacher's "extroversion" and the student's "integrity" alternate. The action is determined to be significant.

As a modification, it is also possible to use, for example, the absolute value of the correlation coefficient instead of the P value in the above-mentioned significance determination. As a specific example, the interfactor significance analysis unit 111 uses the above data (a) to (c) acquired from the personality information storage unit 102 and the performance information storage unit 103 to use the teacher's Big Five factor and the student's Big Five. A correlation analysis process is performed between the product of the Five factors and the test results, and the absolute value | r | of the correlation coefficient is greater than or equal to a predetermined threshold (for example, 0.5) in the combination of the Big Five factors (teacher and student). Determines that "the interaction is significant (and likely)".

The interfactor significance analysis unit 111 is associated with the teacher's Big Five factor, which is determined to have "significant" interaction with each other as a result of processing such as the two-way analysis of variance processing and the correlation analysis processing described above. All combinations of the student's Big Five factors are notified to the model building unit 112 as significant factor combinations.

Returning to the functional block diagram of FIG. 1, in the present embodiment, the model building unit 112 sets an interaction term having only a term related to the combination of factors determined to be significant by the interfactor significance analysis unit 111. , Build a performance estimation model expressed by a multiple regression equation with the set interaction terms.

Here, this performance estimation model will be specifically described later with reference to FIG.
(A) Using performance information (for example, test results) as the objective variable
(B) The factor score (for example, Big Five score) of the first target (for example, a teacher) is set as the first explanatory variable.
(C) The factor score (for example, Big Five score) of the second target (for example, a student) is used as the second explanatory variable.
(D) It is expressed by a multiple regression equation including an interaction term for the first explanatory variable of (b) and the second explanatory variable of (c) above. It should be noted that constructing such a performance estimation model corresponds to determining each coefficient in this multiple regression equation using, for example, the least squares method.

Next, the multiple regression analysis unit 113a of the compatibility score determination unit 113 sets the factor score (Big Five score) of one target (teacher) to be the compatibility determination target (Big Five score) and the other to the interaction term in the constructed performance estimation model. By substituting the factor score (Big Five score) of the target (student), the interaction value, which is the value after the factor score is substituted in this interaction term, is calculated.

The compatibility score determination unit 113 further determines the "compatibility score" related to the compatibility between one object (teacher) and the other object (student) based on the calculated interaction value. In the present embodiment, the compatibility score determination unit 113 further determines the compatibility score. The "compatibility score" for all combinations of the target (teacher) belonging to the target group 1 and the target (student) belonging to the target group 2 is determined.

FIG. 3 is a schematic diagram for explaining an embodiment of the model construction process and the compatibility score determination process in the model construction unit 112 and the compatibility score determination unit 113.

As shown in FIG. 3, the model building unit 112 first of all combinations of the teacher's Big Five factor and the student's Big Five factor, which were determined to be significant by the interfactor significance analysis unit 111. Received notification, then
(A) An interaction term having only a term relating to a combination of these factors, and
(B) A term representing the main effect of the first explanatory variable (teacher's Big Five score) and
(C) Set a multiple regression equation with a term representing the main effect of the second explanatory variable (student's Big Five score).

Next, the model building unit 112
(A) Obtain the Big Five scores of a plurality of teachers (target ID = teacher 01, teacher 02, teacher 03, ...) belonging to the target group 1 from the personality information storage unit 102.
(B) From the personality information storage unit 102, the Big Five scores of a plurality of students (target ID = student 01, student 02, student 03, ...) belonging to the target group 2 are earned.
(C) From the performance information storage unit 103, all the test results of the students of the target group 2 in charge of the teacher of the target group 1, that is, the test results for each combination of the teacher of the target group 1 and the students of the target group 2 are acquired. death,
Using the data of (a) to (c), the parameters of the multiple regression equation set, specifically α ₁ , α ₂ , ···, α ₅ , β ₁ , β in the multiple regression equation of FIG. ₂ , ..., β ₅ , γ ₁ , γ ₂ , ..., γ ₂₅ , are determined by, for example, the least squares method, thereby completing the multiple regression equation as a parameter estimation model.

Here, in the multiple regression equation shown in FIG. 3, for example, the term γ ₁ O ₁ O ₂ is set in the interaction term, which is the teacher's “openness (O ₁ )” and the student's term. It indicates that the combination with "openness (O ₂ )" was notified to be significant. Therefore, in the present embodiment _{, 25 parameters γ n} are not necessarily set, but are set only for the combination notified as significant.

By the way, the teacher's Big Five scores are openness (O ₁ ), integrity (C ₁ ), extroversion (E ₁ ), harmony (A ₁ ), and emotional instability (N ₁ ). There are five student Five scores: openness (O ₂ ), integrity (C ₂ ), extroversion (E ₂ ), harmony (A ₂ ), and emotional instability (N _2). Therefore, the interaction term of the present embodiment includes a maximum of 25 (= 5 × 5) terms.

The model building unit 112 can also set terms related to all combinations of factors, that is, 25 terms, without considering the significance of the interaction terms (in this case, the significance between factors). The sex analysis unit 111 is unnecessary). However, by incorporating only the combination of factors determined to be significant into the interaction term, as described above, it is possible to avoid a situation in which the processing load increases and regression becomes difficult, and more reliably and promptly. The compatibility score can be determined.

Next, the compatibility score determination unit 113 extracts the _{interaction term (γ 1} O ₁ O ₂ + γ ₂ O ₁ C ₂ + ... + γ ₂₅ N ₁ N ₂ ) in the multiple regression equation that is the constructed performance estimation model. Then, the compatibility score Sc is calculated by the following equation (1) Sc ＝ γ ₁ O ₁ O ₂ ＋ γ ₂ O ₁ C ₂ ＋ ・ ・ ・ ＋ γ ₂₅ N ₁ N ₂
In addition to the above equation (1), the factor score of one target (for example, teacher) (for example, Big Five score) and the factor score of the other target (for example, student) (for example, Big Five score) are defined in ) Is substituted to calculate the compatibility score Sc between the objects.

Furthermore, in the present embodiment, the compatibility score determination unit 113 includes a teacher belonging to the target group 1 (target ID = teacher 01, teacher 02, teacher 03, ...) And a student belonging to the target group 2 (target ID = student). 01, Student 02, Student 03, ...) Determines the compatibility score Sc for all combinations.

Returning to the functional block diagram of FIG. 1, the matching unit 114 is based on the compatibility score Sc determined for each combination of the target belonging to the target group 1 (for example, a teacher) and the target belonging to the target group 2 (for example, a student). A target (student) belonging to the second target group that matches the target (teacher) belonging to the first target group is determined.

Here, in the present embodiment, first, the directed bipartite graph generation unit 114a of the matching unit 114
(A) Each object (each teacher) belonging to the first object group is set as the first node.
(B) With each object (each student) belonging to the second object group as the second node
(C) Generate a directed bipartite graph connecting the first node and the second node with an edge.

Next, the matching unit 114
(D) The cost of the edge in (c) above is determined for the combination of the target (teacher) corresponding to the first node related to this edge and the target (student) corresponding to the second node related to this edge. Determined based on the compatibility score Sc, and then
(E) Since the generated directed bipartite graph is solved as the minimum cost flow problem, the target (student) belonging to the second target group that matches the target (teacher) belonging to the first target group is determined. be.

The minimum cost flow problem in (e) above is, for example, non-patent document: Z. Kiraly, P. Kovacs. “Efficient implementation of minimum-cost flow algorithms”, Acta Universitatis Sapientiae, Informatica, vol.4, no. 1 , pp.67-118. It can be solved using the solution algorithm disclosed in 2012.

Further, as a modification mode, the matching unit 114 selects, for example, the objects having the highest compatibility score Sc between the object (teacher) belonging to the first target group and the object (student) belonging to the second target group. It is also possible to sequentially determine as a matching pair (in this case, the directed bipartite graph generation unit 114a becomes unnecessary). However, by performing the process of solving the directed bipartite graph as the minimum cost flow problem as described above, it is possible to more reliably realize perfect matching such that, for example, all students obtain a matching teacher. It becomes.

FIG. 4 is a schematic diagram for explaining an embodiment of the matching process in the matching unit 114.

Here, FIG. 4 shows one target belonging to the second target group (one operator in FIG. 4) that matches one target in the first target group (one customer in FIG. 4). An embodiment of determining the above, that is, an embodiment of performing one-to-one matching is shown. In addition, it is assumed that the compatibility score Sc between these customers and operators is calculated using a performance estimation model that uses the "contract rate" and the "customer satisfaction" obtained from the results of a questionnaire survey of customers as performance information. ..

In the embodiment shown in FIG. 4, the matching unit 114 first
(A) Set the "source" and "sink" so as to sandwich the "directed bipartite graph" generated by the directed bipartite graph generator 114a.
(B) Maximum for each edge in the "directed bipartite graph", including the edge connecting the "source" and each "first node" and the edge connecting each "second node" and the "sink" Set the flow rate to 1.

Here, in this "directed bipartite graph", the cost of each edge corresponds to the target (customer) corresponding to the first node related to the edge and the target (operator) corresponding to the second node related to this edge. ) Is set to -1 times the compatibility score Sc determined for the combination with, that is, -Sc. As a result, the possibility that a larger flow rate is generated increases between the nodes related to the target having the higher compatibility score Sc, and it becomes possible to carry out the matching process so that the compatibility with each other becomes better.

The matching unit 114 applies a solution algorithm for the minimum cost flow problem to the above-mentioned "source" and "sink" and the system of the "directed bipartite graph" sandwiched between them, and is indicated by a thick arrow in FIG. Determine the minimum cost flow, such as. Next, the pair of the target (customer) of the first target group and the target (operator) of the second target group connected by the minimum cost flow (thick line arrow) is output as a matched combination. This output is, as a specific example, in the case of FIG.
((Customer 1a, Operator 2b), (Customer 1b, Operator 2c), (Customer 1c, Operator 2a), ...)
The output will be like this.

In this way, the matching unit 114 determines the optimum target pair (a group consisting of two people, the customer and the operator) from the viewpoint of mutual compatibility, and thereby each target (customer) of the first target group. ), It is possible to arrange the target (operator) of the second target group that is considered to be optimal.

FIG. 5 is a schematic diagram for explaining another embodiment of the matching process in the matching unit 114.

Here, in FIG. 5, N objects belonging to the second object group (N is an integer of 2 or more) matching one object belonging to the first object group (teacher in FIG. 5) (FIG. 5). In 5, an embodiment for determining the student), that is, an embodiment for performing one-to-many matching is shown. In addition, it is assumed that the compatibility score Sc between these teachers and students is calculated using a performance estimation model using the test results of the students who are in charge of the teacher as performance information.

In the embodiment shown in FIG. 5, the matching unit 114 first
(A) Set the "source" and "sink" so as to sandwich the "directed bipartite graph" generated by the directed bipartite graph generator 114a.
(B) Set the maximum flow rate of each edge connecting the "source" and each "first node" to N.

Here, even in this "directed bipartite graph", the cost of each edge corresponds to the target (teacher) corresponding to the first node related to the edge and the target (student) corresponding to the second node related to this edge. ) Is set to -1 times the compatibility score Sc determined for the combination with, that is, -Sc. As a result, the possibility that a larger flow rate is generated increases between the nodes related to the target having the higher compatibility score Sc, and it becomes possible to carry out the matching process so that the compatibility with each other becomes better.

The matching unit 114 applies the solution algorithm of the minimum cost flow problem to the above-mentioned "source" and "sink" and the "directed bipartite graph" system sandwiched between them, and determines the minimum cost flow. Then, a group consisting of one object (teacher) in the first object group and N objects (students) in the second object group connected by the minimum cost flow is output as a matched combination. ..

In this way, the matching unit 114 sets N objects (students) belonging to the second target group that match one target (teacher) belonging to the first target group from the viewpoint of mutual compatibility. It is also possible to determine, for example, the target (student) of the second target group, which should be the personnel of the class in charge of each target (teacher) of the first target group.

Returning to the functional block diagram of FIG. 1, the compatibility score information determined by the compatibility score determination unit 113 (for example, the compatibility score Sc for each combination of target IDs) is stored and managed by, for example, the compatibility score storage unit 104. , Appropriately or according to the user's instruction via the keyboard 106, the display 107 may be displayed via the input / output control unit 122. Further, it is transmitted to an external information processing device, for example, a customer management server owned by a call center via a communication control unit 121 and a communication interface unit 101, and is used there for various purposes, for example, for examining customer response results. preferable.

Further, the matching information determined by the matching unit 114 (for example, a set of target IDs of the personnel forming a group to be in charge of the personnel of each target ID) is also stored and managed by the matching result storage unit 105, for example. , Appropriately or according to the user's instruction via the keyboard 106, the display 107 may be displayed via the input / output control unit 122. Furthermore, it is transmitted to an external information processing device, for example, a business management server owned by an English conversation class, via a communication control unit 121 and a communication interface unit 101, and is used there for various purposes such as determining a glass and a teacher of the class. It is also preferable to be done.

As described in detail above, according to the present invention, it is possible to determine the compatibility between objects by considering the interaction of the characteristics (for example, personality information) of each object. Further, as one application form of the present invention, matching between objects can be performed using the compatibility determined in this way.

This makes it possible to appropriately allocate personnel in various fields in consideration of compatibility between personnel. For example, in providing human services to customers, it is possible to realize more effective or highly satisfying performance. You can also do that. In particular, it is possible to contribute to the creation of an environment that promotes the improvement of students' academic ability, for example, in the field of education.

With respect to the various embodiments of the present invention described above, various changes, modifications and omissions within the scope of the technical idea and viewpoint of the present invention can be easily made by those skilled in the art. The above description is merely an example and does not attempt to limit anything. The present invention is limited only to the scope of claims and their equivalents.

1 Matching device (compatibility determination device)
101 Communication interface unit 102 Personality information storage unit 103 Performance information storage unit 104 Compatibility score storage unit 105 Matching result storage unit 106 Keyboard (KB)
107 Display (DP)
111 Interfactor significance analysis unit 111a Dual distribution analysis unit 112 Model construction unit 113 Compatibility score determination unit 113a Multiple regression analysis unit 114 Matching unit 114a Directed bipartite graph generation unit 121 Communication control unit 122 Input / output control unit 2 Personality information Database (DB)
3 Performance information DB

Claims (9)

A compatibility determination program that operates a computer that determines compatibility information between subjects having a predetermined characteristic expressed by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. The factor score of one target and the factor score of the other target for which the compatibility is determined with respect to the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the target. A compatibility determination program characterized in that a computer functions as a compatibility score determining means for determining a compatibility score relating to compatibility between one object and the other object based on a value obtained by applying the above. The first target group is used by using the result value obtained for each combination of each target of the first target group to which the one target belongs and each target of the second target group to which the other target belongs. For each combination of each factor score of the subject belonging to the second target group and each factor score of the subject belonging to the second target group, a value indicating the significance of the interaction between the factor scores related to the combination is calculated. Sex analysis means and
Model construction that determines an interaction part having only a term for the factor score related to the combination judged to be significant based on the value indicating the significance, and constructs the model having the determined interaction part. The compatibility determination program according to claim 1, wherein the computer is further functioned as a means. The subject is a person, the predetermined trait is a personality trait, and the resulting value is either the first subject or the second subject alone or jointly. The compatibility determination program according to claim 1 or 2, wherein the score, rank, or rank is given to the result of a predetermined action. A matching program that functions a computer that determines matching between objects belonging to each group in a first target group and a second target group, which are a set of objects having a predetermined characteristic represented by at least one factor score. And
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. For the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the object, each factor score of the object belonging to the first object group and the second object Based on the value obtained by applying the combination with each factor score of the subject belonging to the group, for each combination of each subject belonging to the first target group and each subject belonging to the second target group, the combination of the combination A compatibility score determining means for determining compatibility scores related to compatibility, and
Based on the compatibility score determined for each combination, the computer functions as a matching means for determining the target belonging to the second target group, which matches the target belonging to the first target group. Matching program. In the matching means, each target belonging to the first target group is set as a first node, each target belonging to the second target group is set as a second node, and the first node and the second node are edges. A connected bipartite graph was generated and the cost of the edge was determined for the combination of the object corresponding to the first node for the edge and the object corresponding to the second node for the edge. After making a decision based on the compatibility score, the bipartite graph is solved as the minimum cost flow problem, and the target belonging to the second target group that matches the target belonging to the first target group is determined. The matching program according to claim 4. The matching means has a maximum flow rate of 1 for each edge according to the bipartite graph, including an edge connecting the set source and the first node and an edge connecting the second node and the set sink. The matching according to claim 5, wherein one object belonging to the second object group is determined by solving the minimum cost flow problem to match one object belonging to the first object group. program. The matching means belongs to the first target group by solving the minimum cost flow problem with the maximum flow rate of each edge connecting the set source and the first node as N (N is an integer of 2 or more). The matching program according to claim 5, wherein N objects belonging to the second target group that match one target are determined. A compatibility determination device that determines compatibility information between objects having a predetermined characteristic expressed by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. The factor score of one target and the factor score of the other target for which the compatibility is determined with respect to the interaction part in the constructed model provided with the interaction part with respect to the second explanatory variable which is the factor score of the target. A compatibility determination device comprising a compatibility score determining means for determining a compatibility score relating to compatibility between the one object and the other object based on a value obtained by applying the above. A computer-based compatibility determination method for determining compatibility information between subjects having a predetermined characteristic expressed by at least one factor score.
A model in which the result value, which is a value related to the result of the action of the first object and the second object, is used as the objective variable, the first explanatory variable, which is the factor score of the first object, and the second. Steps to build a model with an interaction with the second explanatory variable, which is the factor score of interest,
The one object and the other object are based on the values obtained by applying the factor score of one object and the factor score of the other object, which are the compatibility determination targets, to the interaction part in the constructed model. A compatibility determination method comprising a step of determining a compatibility score relating to compatibility with.

Images