A system and method for supporting behavioral change in target individuals.

Abstract

Support the behavioral change of the individual in accordance with the characteristics of the environment / group to which the individual belongs. [Solution] The system integrates data on the behavior of each of several subjects and data on the characteristics of the environment / group to which the subject belongs, calculates a cluster index value using this data, and classifies the subject into one of several clusters moving toward the target behavior, in which the cluster index value calculated for the subject belongs is associated with that cluster. For each of the multiple clusters, the system selects a policy selection method from among several policy selection methods that is suitable for the cluster characteristics, which are the characteristics of the environment / group to which the subject classified into that cluster belongs, and selects a policy for the subject classified into that cluster according to the selected policy selection method.

Description

Technical Field

[0001] The present invention relates to a technique for assisting in the behavioral transformation of a target person using a computer.

Background Art

[0002] Regarding this type of technology, a patent invention of the applicant of the present application is known (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Evaluations of a target person (for example, to which cluster the target person is classified and measures taken for the target person toward the intended behavioral transformation) vary depending on the evaluation perspective of the target person.

[0005]

Means for Solving the Problems

[0006] The system integrates data on the behavior of each of the multiple subjects and data on the characteristics of the environment / group to which the subject belongs. Using this data, it calculates a cluster index value and classifies the subject into one of several clusters moving towards the target behavior, based on the cluster to which the calculated cluster index value belongs. For each of the multiple clusters, the system selects a policy selection method from among several policy selection methods that is suitable for the cluster characteristics, which are the characteristics of the environment / group to which the subject classified into that cluster belongs. According to the selected policy selection method, the system selects policies for the subject classified into that cluster. [Effects of the Invention]

[0007] It is expected that this will support behavioral change in the target individual in accordance with the characteristics of the environment / group to which the target individual belongs. [Brief explanation of the drawing]

[0008] [Figure 1] An overview of the embodiment is shown. [Figure 2] An example of the overall system configuration according to the embodiment is shown. [Figure 3] An example of a characteristics database configuration is shown. [Figure 4] An example of the structure of integrated data is shown. [Figure 5] An example of a matching database configuration is shown. [Figure 6] An example of policy selection following a correlation method is shown. [Figure 7] An example of policy selection following the before-and-after method is shown. [Modes for carrying out the invention]

[0009] In the following explanation, "interface device" may refer to one or more interface devices. These one or more interface devices may be at least one of the following: • One or more I / O (Input / Output) interface devices. An I / O (Input / Output) interface device is an interface device to at least one of the following: an I / O device and a remote display computer. The I / O interface device to the display computer may be a communication interface device. The at least one I / O device may be either a user interface device, such as an input device like a keyboard and a pointing device, or an output device like a display device. • One or more communication interface devices. One or more communication interface devices may be one or more identical communication interface devices (e.g., one or more NICs (Network Interface Cards)) or two or more different communication interface devices (e.g., a NIC and an HBA (Host Bus Adapter)).

[0010] Furthermore, in the following explanation, "memory" refers to memory separate from the NVM in the NVM drive, and consists of one or more memory devices, which are typically main memory devices. At least one memory device in the memory may be a volatile memory device or a non-volatile memory device.

[0011] Furthermore, in the following explanation, "persistent storage device" refers to one or more persistent storage devices. Persistent storage devices are typically non-volatile storage devices (e.g., auxiliary storage devices), specifically, for example, HDDs (Hard Disk Drives) or SSDs (Solid State Drives).

[0012] Furthermore, in the following explanation, "storage device" may refer to at least memory, including both memory and persistent storage.

[0013] Furthermore, in the following explanation, "processor" refers to one or more processor devices. At least one processor device is typically a microprocessor device such as a CPU (Central Processing Unit), but may be other types of processor devices such as a GPU (Graphics Processing Unit). At least one processor device may be single-core or multi-core. At least one processor device may be a processor core. At least one processor device may be a broader processor device such as a hardware circuit that performs some or all of the processing (e.g., an FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)).

[0014] Furthermore, in the following explanation, "DB" is an abbreviation for database. Also, the expression "xxx table" is sometimes used to describe information from which an output is obtained for an input, but this information can be data with any structure, or a learning model such as a neural network that generates an output for an input.Therefore, "xxx table" can be called "xxx information".In addition, in the following explanation, the structure of each table is just an example, and one table may be divided into two or more tables, or all or part of two or more tables may be a single table.

[0015] Furthermore, in the following explanation, functions may be described using the expression "kkk section," but functions may be implemented by the execution of one or more computer programs by a processor, or by one or more hardware circuits (e.g., FPGA or ASIC). When a function is implemented by the execution of a program by a processor, the defined processing is carried out using memory and / or interface devices as appropriate, so the function may be at least a part of the processor. Processing described with a function as the subject may be processing performed by the processor or a device having that processor. Programs may be installed from program source. Program source may be, for example, a program distribution computer or a computer-readable recording medium (e.g., a non-temporary recording medium). The description of each function is an example, and multiple functions may be combined into one function, or one function may be divided into multiple functions.

[0016] Furthermore, in the following explanation, a system that supports behavioral change in a target individual will be referred to as a "support system." A "support system" may be one or more physical computers, a software-defined system implemented by at least one physical computer running predetermined software, or a system implemented on a cloud infrastructure (typically a combination of multiple types of computing resources, including processors and memory devices). For example, if a computer has a display device and displays information on its own display device, that computer may be a support system. Also, for example, if a first computer (e.g., a server) transmits output information to a remote second computer (a display computer (e.g., an administrator terminal described later)) and the display computer displays that information (the first computer displays information on the second computer), then at least the first computer of the two computers may be a support system. In other words, "displaying output information" by a support system may mean displaying the output information on a display device owned by the computer, or the computer transmitting the output information to the display computer (in the latter case, the output information is displayed by the display computer).

[0017] In the following description, when describing elements of the same type without distinction, the common part of the reference signs is used, and when describing elements of the same type separately, reference signs may be used. For example, when describing clusters without distinction, it may be referred to as "cluster 50", and when describing clusters separately, it may be referred to as "cluster 50A", "cluster 50B", etc.

[0018] Hereinafter, one embodiment will be described. In the following embodiment, the goal of causing behavioral variation for each target person is to have more target persons purchase product X, and more preferably, to have more target persons become repeat purchasers of product X.

[0019] FIG. 1 shows an overview of the embodiment.

[0020] The support system 100 includes a data integration unit 110, a clustering unit 115, a policy selection unit 120, and a process execution unit 140. Also, as data, there are behavior data 150 including various data measured for the behaviors of a plurality of target persons, and related data 160 which is at least a part of the data other than the behavior data 150. The behavior data 150 includes quantitative data 151 and qualitative data 152. The related data 160 includes a matching DB 161, a characteristic DB 162, and a history table 163. The matching DB 161 is a DB referred to for policy selection and includes data in which inhibition factors, policy selection methods, and policies are associated. The characteristic DB 162 includes data conforming to the characteristics of the environment / group (environment and / or group). The history table 163 represents the history of past performance of policy execution.

[0021] The data integration unit 110 integrates, for each subject, subject behavior data, which is data about the subject from the behavioral data, and subject characteristic data, which is data about the characteristics of the environment / group to which the subject belongs from the characteristics DB 162. The data integration unit 110 generates integrated data 10 that includes the integrated data for each subject. For each subject, the data integration unit 110 calculates a total point used for clustering using the subject behavior data and the subject characteristic data. Specifically, for example, qualitative data from the subject behavior data and subject characteristic data are quantified (e.g., as coefficients), and the data integration unit 110 records the calculated total point for each subject in the integrated data 10. The total point is an example of a cluster index value, which is an index value used for clustering. Note that the environment / group to which a subject belongs may be identified from the environment / group (e.g., community) corresponding to the source of data about the subject, or it may be identified from the profile data of the subject, which may contain data representing the environment / group to which the subject belongs.

[0022] There are multiple clusters 50 (for example, 50A to 50E). A cluster axis 55 may also be provided. The cluster axis 55 is an axis that serves as a criterion for the multiple clusters 50 moving toward the target behavior, and in this embodiment, it is an axis corresponding to the total points. For example, by defining the range from the minimum to the maximum value of the total points, a point range (an example of an index value range) for each cluster 50 can be obtained. The point ranges do not overlap between clusters 50. The clustering unit 115 identifies the total points recorded in the integrated data 10 for each subject and classifies the subject into the cluster 50 to which the total points belong. The point ranges between clusters 50 may be uniform or may differ depending on the cluster 50.

[0023] If the correlation method is selected, behavioral changes 60, inhibiting factors 70, and measures 80 are identified or selected in relation to the desired effect for cluster 50. If the before-and-after method is selected, behavioral changes 60, inhibiting factors 70, and measures 80 are identified or selected in relation to the gap between the behavior represented by the subject behavior data of subjects classified into a cluster and the behavior represented by the subject behavior data of subjects classified into the next cluster with the highest point range.

[0024] The processing execution unit 140 performs result processing (for example, displaying or executing the selected measures 80) for each cluster. The processing execution unit 140 also adds past performance, which is an evaluation of the execution results of measures 80, to the history represented by the history table 163. In other words, the history table 163 is updated.

[0025] According to support system 100, for each target individual, target individual behavior data and target individual characteristics data are integrated, and a total score is calculated using this integrated data. Therefore, for each target individual, the total score used for clustering reflects the characteristics of the environment / group to which the target individual belongs. Furthermore, the policy selection method used when selecting policies is a method suitable for the cluster characteristics, which are the characteristics of the environment / group to which the target individuals classified into clusters belong. In this way, the characteristics of the environment / group are reflected in both clustering and policy selection for each cluster. This is expected to support behavioral change in target individuals according to the characteristics of the environment / group to which they belong.

[0026] This embodiment will now be described in detail.

[0027] Figure 2 shows an example of the overall system configuration according to the embodiment.

[0028] A client system 240, a storage server 250, an administrator terminal 210, and a support server 200 are connected to a communication network (e.g., the Internet) 290. Of the storage server 250, administrator terminal 210, and support server 200, at least the support server 200 is a component of the support system 100.

[0029] The client system 240 includes one or more client devices 241. The client devices 241 may be client terminals 241A, such as personal computers, or various sensor devices 241B. The client system 240 transmits at least one type of data measured for multiple subjects to the storage server 250. This at least one type of data, or processed data thereof, constitutes at least a portion of the behavioral data 150.

[0030] The storage server 250 stores the behavioral data 150. The behavioral data 150 includes activity data 151A as an example of quantitative data 151, and profile data 152A and subject survey data 152B as an example of qualitative data 152. Activity data 151A is quantitative data measured for various activities of multiple subjects, and may include, for example, human body measurements, sensor data, machine data, PC operation logs, communication information such as email and SNS (Social Networking Service), information on images (still images or videos) taken by cameras such as security cameras, historical information such as purchase history and work history, management information created for business management, work schedule information, voice data, location identification information (for example, information indicating a location identified using infrared, Wi-Fi (registered trademark), UWB (Ultra Wide Band) (registered trademark), etc.). Activity data 151A may include two or more types of quantitative data 151. Profile data 152A includes a dataset showing the profile of each of several subjects (e.g., age, gender, etc.). Subject survey data 152B includes a dataset showing the results of a survey (e.g., multiple actions) for each of several subjects (e.g., performance for each of multiple actions), such as a dataset showing the results of a questionnaire (multiple questions) (answers to each of multiple questions), or a dataset showing the results of an interview (multiple questions) (answers to each of multiple questions). Storage server 250 may be a server as a digital marketing tool, specifically, for example, a server that scores subjects (e.g., customers or prospective customers). Quantitative data may include a score (e.g., a numerical value) for each subject for each of one or more items.

[0031] The administrator terminal 210 may be an information processing terminal (e.g., a personal computer) that functions as a client (e.g., an input / output console) of the support server 200. The administrator terminal 210 is used by the administrator. The "administrator" may be the administrator of the support server 200, or an administrator who manages multiple target users. The administrator may also be one of the target users.

[0032] The support server 200 includes an interface device 261, a storage device 262, and a processor 263 connected thereto. The interface device 261 is connected to a communication network 290.

[0033] The storage device 262 stores, for example, related data 160 and one or more programs (not shown) that are executed by the processor 263 to realize the data integration unit 110, clustering unit 115, policy selection unit 120, and processing execution unit 140. The related data 160 includes the matching DB 161, characteristics DB 162, and history table 163 described above. The storage device 262 stores the integrated data 10 generated by the data integration unit 110. At least a portion of the related data 160 (for example, at least a portion of the characteristics DB 162) may be stored in a server as a communication interface tool, specifically, a server that provides a service (for example, an SNS (Social Networking Service)) that enables two-way communication between target individuals within a community, which is an example of an environment / group.

[0034] For example, the processor 263 of the support system 100 may perform the following: The processor 263 of the support system 100 acquires customer management data for digital marketing (for example, data representing actions based on various business and management systems such as purchasing behavior, work behavior, learning behavior, application behavior, and social support behavior) and community data (for example, user action data and user environment information within the community). If qualitative research such as interviews is being conducted, the processor 263 also acquires that qualitative data. Processor 263 uses quantitative data 151 for at least a portion of customer management data, quantitative data 151 for at least a portion of individual behavioral data within the community, qualitative data 152 for at least a portion of qualitative personal information data obtained through qualitative research, characteristics DB 162 for at least a portion of user environment information and group characteristics data for the entire community, and history table 163 for at least a portion of data on past policy effects and implemented policies. The processor 263 (e.g., the data integration unit 110) calculates the total points of individuals belonging to the community based on five variable quantities that are likely to correlate with purchasing behavior (an example of a purpose) (e.g., browsing behavior, engagement behavior, the customer's expected budget, lifecycle timing, and the amount of specific actions that contribute to purchasing (e.g., requesting a quote)), and classifies the individuals into clusters 50. The processor 263 (e.g., the clustering unit 115) sets an objective for each cluster, specifying what kind of behavior is desirable. The objective may include behavioral change, and may further include behavioral goals for behavioral change. The behavioral goals may be determined based on the gap between the cluster and the next highest cluster (e.g., the technology disclosed in Patent Document 1), or they may be determined as characteristic behaviors of individuals classified into the cluster (e.g., behaviors that contribute to sales). The "characteristic behaviors of individuals classified into the cluster" may be identified based on past performance corresponding to the cluster characteristics (past performance in the history represented by the history table 163). For example, for each past performance, the past performance may include the cluster characteristics and the behaviors of individuals classified into the cluster having those cluster characteristics. The processor 263 (for example, the policy selection unit 120) performs a matching of previously implemented policies aimed at the objective, including the action goal, identified from the history table 163, with policies for the action to be promoted, as a requirement definition for achieving the action goal. The matching can be based on the similarity between the effect classification of each policy (for example, the policy matching table contains the actual results of policies such as browsing and visit effects) and the expected effect of the action goal (for example, the action that is expected to be browsing in order to increase browsing behavior), and policies with a relatively high degree of similarity may be selected. The processor 263 (e.g., the processing execution unit) acquires the actions of the target cluster after the implementation of the measure and adds the acquired actions, including past performance, to the history represented by the history table 163. Past performance may include the results of a comparison with the period before the implementation of the measure. Past performance may also include results measured inside or outside the support system 100 (e.g., actions such as each target person classified into a cluster viewing the web page for product X). Past performance may also include the results of manual surveys (e.g., survey results represented by data entered from the administrator terminal 210).

[0035] Figure 3 shows an example configuration of the characteristics DB162.

[0036] The characteristics database 162 includes an environmental characteristics table 301 and a population characteristics table 302.

[0037] The environmental characteristics table 301 has an entry for each environmental characteristic. Each entry holds data such as an environmental characteristic ID 311, an environmental characteristic name 312, and an environmental characteristic impact 313. The environmental characteristic ID 311 represents the ID of the environmental characteristic. The environmental characteristic name 312 represents the name of the environmental characteristic (e.g., a name that describes its content). The environmental characteristic impact 313 represents the impact of the environmental characteristic, for example, at least one of the following: (a) a value used to calculate the overall score (e.g., a coefficient), (b) a value that affects the correlation method, (c) a value that affects the before-and-after method, and (d) an impact on the identification or selection of at least one of behavioral changes, inhibitors, and measures (e.g., the degree of impact on each behavioral change, each inhibitor, or each measure). In addition, for each environmental characteristic, the environmental characteristic impact 313 may include subject-related information associated with that environmental characteristic, and subject-related information may include information representing the subject's profile or the group to which the subject belongs (e.g., a community or organization). The data integration unit 110 searches the columns of the environmental characteristics impact 313 for each subject using the subject-related information of that subject as a key to identify the environmental characteristics entry corresponding to that subject, and then adds the environmental characteristics 404A (numerical value) described later to the integrated data 10 using the data contained in that entry.

[0038] The group characteristics table 302 has an entry for each group characteristic. Each entry holds data such as a group characteristic ID 321, a group characteristic name 322, and a group characteristic impact 323. The group characteristic ID 321 represents the ID of the group characteristic. The group characteristic name 322 represents the name of the group characteristic (e.g., a name that describes its content). The group characteristic impact 323 represents the impact of the group characteristic, for example, at least one of the following: (a) a value used to calculate the overall score (e.g., a coefficient), (b) a value that affects the correlation method, (c) a value that affects the before-and-after method, and (d) an impact on the identification or selection of at least one of behavioral changes, inhibitors, and measures (e.g., the degree of impact on each behavioral change, inhibitor, or measure). In addition, for each group characteristic, the group characteristic impact 323 may include the aforementioned subject-related information linked to that group characteristic. The data integration unit 110 searches the columns of the group characteristics influence 323 for each subject using the subject-related information of that subject as a key to identify the group characteristics entry corresponding to that subject, and then adds the group characteristics 404B (numerical value) described later to the integrated data 10 using the data contained in that entry.

[0039] Here, examples of the influence of environmental / group characteristics, such as (a) through (d) above, are as follows: For example, the total score is a good example of a dependent variable, while one or more numerical values ​​obtained from quantitative data on the subject's behavior, one or more numerical values ​​determined based on qualitative data on the subject's behavior (e.g., a coefficient for a certain numerical value obtained from quantitative data), or one or more numerical values ​​determined based on (a) corresponding to the environmental / group characteristics identified from the subject identification data (e.g., a coefficient for a certain numerical value obtained from quantitative data) are good examples of independent variables. According to (b) and (c), for each cluster, the method with the larger value, obtained by comparing (b) and (c) corresponding to the cluster characteristics, may be adopted as the policy selection method corresponding to the cluster. Note that this is just one example of a method for selecting a policy selection method according to cluster characteristics. The policy selection method may also be selected based on the features of the cluster characteristics. According to (d), for each cluster, at least one of the following—behavioral change, inhibiting factors, and measures—is identified or selected based on the characteristics of the cluster.

[0040] Furthermore, the characteristics of the environment / group may be obtained from publicly available information (e.g., research papers) or defined based on publicly available information.

[0041] Figure 4 shows an example of the configuration of the integrated data 10.

[0042] The integrated data 10 has an entry for each subject. The entry includes the subject's ID 401, numerical values ​​402 for each behavior represented by quantitative data from the subject's behavior data (e.g., 402A, 402B, ...), numerical values ​​403 set based on qualitative data from the subject's behavior data (e.g., 403A, 403B, ...), numerical values ​​404 set based on environmental / group characteristics obtained from subject characteristic data (404A, 404B, ...), and the calculated and recorded total points 405. Each of the numerical values ​​402 to 404 may be prepared for each item. The numerical value 402 may be a value obtained from customer lead or member data obtainable from digital marketing tools or community tools, and may be a value as a behavior for each item such as web page views, email newsletter opens, exhibition participation, or "like" reactions (e.g., the number of times the subject performed the action). The numerical value 403 may be a value based on interviews or surveys with people or organizations, sales experience, community manager knowledge, etc.

[0043] The integrated data 10 is generated by the data integration unit 110. Specifically, for each of the multiple subjects, the subject behavior data includes quantitative and qualitative data, and the data integration unit 110 calculates the subject's total points based on the numerical value 402 represented by the quantitative data for each of the subject's one or more items, the numerical value 403 set based on the subject's qualitative data (for example, a coefficient for any of the numerical values ​​402), and the numerical value 404 set based on the characteristics of the environment / group to which the subject belongs (for example, a coefficient for any of the numerical values ​​402). The total points can also be described as the expected value of achieving the goal (purchasing behavior). For each explanatory variable item of the total points (expected value), a value identified for that item from the quantitative data of the subject behavior data can be adopted as the explanatory variable, and a value determined based on the qualitative data of the subject behavior data or a value determined based on the subject characteristics data can be adopted as the coefficient (weight) of the explanatory variable.

[0044] Figure 5 shows an example configuration of the matching DB161.

[0045] Matching DB161 may include index 501 and policy matching table 502.

[0046] Index 501 contains key information 511 and reference information 512 for each entry. Key information 511 is information (e.g., a hash) based on at least one of the following: cluster characteristics and selected policy selection method. Reference information 512 is information representing the reference in the policy matching table 502.

[0047] The policy matching table 502 contains data for each entry, including factor ID 521, inhibiting factor 522, method 523, behavioral change 524, behavioral goal 525, policy ID 526, and policy 527.

[0048] Factor ID 521 represents the ID of the inhibiting factor. Inhibiting factor 522 represents the inhibiting factor.

[0049] Method 523 represents the policy selection method. Behavioral change 524 represents behavioral change. Behavioral goal 525 represents the behavioral goal. Policy ID 526 represents the policy ID. Policy 527 represents the policy. Policy 527 may include supplementary data such as the characteristics of the policy in addition to the policy to be implemented.

[0050] Figure 6 shows an example of policy selection following a correlation method.

[0051] For example, suppose a correlation method is selected based on the cluster characteristics of cluster 50Q among clusters 50P to 50R. The correlation method is a method of selecting measures that have a relatively high correlation with the behaviors represented by the behavioral data of the target persons classified in cluster 50Q for the purpose corresponding to cluster 50Q. The measure selection unit 120 selects the correlation method if the cluster characteristics of cluster 50Q are defined as cluster characteristics that are suitable for overall transition.

[0052] Here, "cluster characteristics" refer to the characteristics of the environment / group to which the subjects classified into a cluster belong, as described above. The cluster characteristics may be the most prevalent environmental / group characteristics among the environmental / group characteristics of all subjects classified into a cluster, or among the environmental / group characteristics of subjects who meet a predetermined condition (for example, subjects whose behavioral data shows behaviors that meet a predetermined condition).

[0053] Furthermore, the "purpose corresponding to cluster 50Q" may be a predetermined purpose for cluster 50Q, or it may be a purpose determined based on characteristics obtained from the behavioral data and characteristic data of all subjects classified into cluster 50Q. In either case, the purpose of cluster 50Q is to move towards (or approach) the target behavior. The "target behavior" can be the purchasing behavior of product X.

[0054] Furthermore, the "behavior represented by the subject behavior data of subjects classified into cluster 50Q" may be the most frequent behavior (e.g., a common behavior) among the behaviors of all subjects classified into the cluster, or among the behaviors of subjects that meet the specified conditions.

[0055] Furthermore, "measures that have a relatively high correlation with the behaviors represented by the behavioral data of individuals classified in Cluster 50Q for the purpose corresponding to Cluster 50Q" can be any measure that contributes to the purpose corresponding to Cluster 50Q and has a relatively high correlation (for example, the highest) with the behaviors represented by the behavioral data of individuals classified in Cluster 50Q. Information on which purpose each measure contributes to may be included in the measure 527 for each measure in the measure matching table 502. In addition, the correlation between behavior and measures may be identified from the history represented in the history table 163 (for example, the behaviors, implemented measures, and measures effects included in each past performance), or the measure 527 for each measure may include information representing the behaviors defined as having a high correlation with that measure.

[0056] Furthermore, "overall transition" refers to fostering the behavior of the entire cluster toward a target behavior, specifically enabling cluster 50Q as a whole to transition to cluster 50R or a higher cluster 50. "Cluster characteristics defined as suitable for overall transition" can be cooperative or similar environmental / group characteristics.

[0057] A specific example of the correlation method is shown in Figure 6.

[0058] In other words, let's assume that the cluster characteristic of cluster 50Q is "high awareness of other members." This characteristic means that it includes many organizations and individuals where members belonging to the same environment / group understand each other well and are in a cooperative relationship. Because their state of mind (e.g., emotions) is stable, they get angry or sad less often and tend to view things positively.

[0059] In this case, the correlation method is determined from the cluster characteristics of cluster 50Q. Specifically, for example, the correlation method is identified from the environmental characteristic influence 313 and / or group characteristic influence 323 of the cluster characteristics of cluster 50Q. Then, from the key information 511 that matches the information based on the cluster characteristics and correlation method of cluster 50Q, the entry with factor ID "0002" is identified as the reference destination represented by the reference destination information 512. Since the entry with factor ID "0002" is for the correlation method, the measures for the before-and-after method are left blank. Furthermore, for cluster 50Q in which the correlation method has been selected, in order to reduce the inhibiting factors corresponding to the cluster characteristics of cluster 50Q and realize behavioral changes corresponding to the said cluster characteristics and correlation method, there may be more than one pair of behavioral change and measures defined. For example, in order to increase the average overall score of cluster 50Q, measures for behavioral change "maintaining the status quo" and measures for behavioral change "raising the stage" (transition to a higher cluster) are selected.

[0060] Figure 7 shows an example of policy selection following the before-and-after method.

[0061] For example, suppose the before-and-after method is selected based on the cluster characteristics of cluster 50K among clusters 50J to 50L. The before-and-after method is a method of selecting measures that correspond to the gap between the characteristics of cluster 50K and the characteristics of cluster 50L, which has the next highest point range after cluster 50K. The measure selection unit 120 selects the before-and-after method if the cluster characteristics of cluster 50K are defined as cluster characteristics suitable for individual transitions.

[0062] Here, "cluster characteristics" can refer to feature quantities or similar information related to the cluster, for example, the behaviors represented by the behavioral data of subjects classified into that cluster. An example of behavior as a cluster characteristic may be a behavior obtained from the statistics of the behaviors of all subjects belonging to that cluster (for example, the behavior taken by the most subjects), or it may be the behavior of a representative subject.

[0063] Furthermore, an example of a "gap" in characteristics between clusters is a set of item values ​​(elements) that differ among the 50 clusters, among multiple item values ​​(e.g., instances representing behavior) of one or more items related to cluster characteristics.

[0064] Furthermore, "measures addressing gaps" are measures that are expected to bridge the gap between the characteristics of the lower cluster and the characteristics of the higher cluster; in other words, measures that are expected to make the gaps in the characteristics of the lower cluster match the characteristics of the higher cluster. For each measure, information on what kind of gap it contributes to bridging (for example, the set of item values ​​that make up the gap or the characteristics of the gap) may be included in the measure 527 for each measure in the measure matching table 502.

[0065] Furthermore, "individual transition" refers to moving as many participants as possible from one cluster to the next higher cluster, specifically, enabling as many participants as possible to move from cluster 50K to cluster 50L. "Cluster characteristics defined as suitable for individual transition" can be those of an environment / group that lacks or has similar characteristics to lack cooperation.

[0066] A concrete example of the front-to-back method is shown in Figure 7.

[0067] In other words, let's assume that the cluster characteristic of cluster 50K is "low awareness of other members." This characteristic means that members belonging to the same environment / group do not understand each other, and that it includes many individuals who are not in an environment / group where they can cooperate. Their state (e.g., emotions) becomes unstable, and their attitudes, behaviors, and judgments change significantly from day to day.

[0068] In this case, the before-and-after method is determined from the cluster characteristics of cluster 50K. Specifically, for example, the before-and-after method is identified from the environmental characteristic influence 313 and / or group characteristic influence 323 of the cluster characteristics of cluster 50K. Then, from the key information 511 that matches the information based on the cluster characteristics of cluster 50K, the before-and-after method, and the gap between cluster 50K and cluster 50L, the entry with factor ID "0001" is identified as the reference destination represented by the reference destination information 512. Since the entry with factor ID "0001" is for the before-and-after method, the measures for the correlation method are left blank. Furthermore, for cluster 50K in which the before-and-after method is selected, only one pair of behavioral change and measures is needed in order to reduce the inhibiting factors corresponding to the gap between cluster 50K and the next cluster 50L and to realize behavioral change corresponding to the cluster characteristics and correlation method.

[0069] In addition to the correlation method, the key information 511 may also be information that is compatible with the cluster characteristics of cluster 50Q, as well as the correlation method.

[0070] Regardless of whether the policy selection method is correlation-based or before-and-after, the following terminology definitions may be adopted. "Inhibiting factors" are factors that hinder behavioral changes necessary to move from a lower cluster to a higher cluster (one level higher or higher). "Cognitive bias" is a broader concept than "inhibiting factors," and is a bias defined based on findings from behavioral science. Cognitive bias is the same as, or even broader than, so-called "nudges." Therefore, inhibiting factors can be considered a concept encompassed within cognitive bias.

[0071] The above is a description of this embodiment. A summary or supplementary explanation of this embodiment is as follows. The following explanation may include modifications of this embodiment.

[0072] The support system 100 comprises a data integration unit 110, a clustering unit 115, a policy selection unit 120, and a processing execution unit 140. The data integration unit 110 integrates, for each target person, the target person behavior data from the behavior data 150 for that target person and the target person characteristic data from the characteristic DB 162 regarding the characteristics of the environment / group to which the target person belongs, and calculates a total score using the target person behavior data and the target person characteristic data. The clustering unit 115 classifies the target person into one of several clusters 50, to which the point range to which the calculated total score for that target person belongs is associated. The policy selection unit 120 selects a policy selection method suitable for the cluster characteristics of each cluster 50, and selects policies for the target people classified into that cluster according to the selected policy selection method. The processing execution unit 140 executes result processing, which is processing related to the selected policies, for each cluster 50.

[0073] An example of the technical significance of data integration is improved clearing accuracy. Specifically, for each subject, subject behavior data is associated with subject characteristic data representing the characteristics of the environment / group to which the subject belongs (e.g., an organization or a community with similar interests) through data integration. Subject characteristic data is typically qualitative data, but it can also be quantitative data. When integrating this data, for the qualitative data within the integrated data, a value is determined based on what the qualitative data represents, and this determined value is set in the integrated data. The quantification of subject behavior data and qualitative data within subject characteristic data may be done manually or automatically using methods such as a model (e.g., a trained model that takes qualitative data as input and outputs numerical values). By adding environmental / group characteristics as coefficients (or other numerical values) in the calculation of the overall points used for clustering, each subject can be classified into a cluster with higher accuracy.

[0074] Multiple policy selection methods may include correlation methods and prior-to-priority methods. For each cluster, the choice between the correlation method and the prior-to-priority method may depend on the cluster characteristics (environmental / group characteristics) of that cluster. As a method for selecting a policy selection method based on cluster characteristics, for example, the following methods can be adopted.

[0075] In other words, as described above, in the characteristics DB162, there is an environmental characteristic influence 313 for each environmental characteristic and a group characteristic influence 323 for each group characteristic. Each of these influences 313 and 323 includes information used as a basis for determining which policy selection method to choose, as well as the subject-related information described above. For each subject, an environmental characteristic influence 313 and / or group characteristic influence 323 including the subject-related information for that subject are identified, and for each cluster, the cluster characteristics are determined from the environmental / group characteristics corresponding to the subjects classified in that cluster. The policy selection method can be selected using the influences 313 and / or 323 corresponding to the cluster characteristics. Furthermore, influences 313 and / or 323 include information set so that the correlation method is selected if the influence 313 or 323 corresponds to an environmental / group characteristic defined as suitable for overall transition (for example, a value "100" that influences the correlation method and a value "0" that is provided for the before-and-after method). On the other hand, influences 313 and / or 323 include information (for example, a value "0" that affects the correlation method and a value "100" that provides to the correlation method) that is set to select the before-and-after method if the influences 313 and 323 correspond to environmental / group characteristics defined as being suitable for individual transitions. As a result, for each cluster 50, the policy selection unit 120 selects the correlation method for clusters with cluster characteristics defined as being suitable for overall transitions, and selects the before-and-after method for cluster characteristics defined as being suitable for individual transitions. Note that each of influences 313 and 323 may include information indicating whether the environmental / group characteristics corresponding to that influence are defined as characteristics suitable for overall transitions or characteristics defined as being suitable for individual transitions (or information that serves as a basis for determining whether it is suitable for overall transitions or individual transitions).

[0076] The processing execution unit 140 identifies key information 111 for each cluster 50 that matches the information containing information representing the cluster characteristics and the selected policy selection method, and can select a policy from the reference destination represented by the reference destination information 112 corresponding to the identified key information 111. "Key information 111 that matches the information containing information representing the cluster characteristics and the selected policy selection method" may be, for example, key information 111 that matches the feature quantities of the information containing information representing the cluster characteristics and the selected policy selection method. The information containing information representing the cluster characteristics and the selected policy selection method may further include information representing cluster characteristics, or other information. Furthermore, in the above embodiment, when selecting a policy based on the cluster characteristics and the policy selection method, if the policy selection method is a correlation method, factors hindering the achievement of the cluster's objective and behavioral changes for achieving the objective are also identified, and if the policy selection method is a pre- and post-method method, factors hindering gap overcoming and behavioral changes for achieving gap overcoming are also identified. The identification of inhibiting factors, the identification of behavioral changes, and the selection of measures may be carried out in parallel (for example, simultaneously) or sequentially (for example, behavioral changes and inhibiting factors may be identified first, and then measures may be selected).

[0077] The policy selection unit 120, when selecting a policy for each cluster 50 from the references represented by the identified reference information 122, may select a policy for that cluster based on the past performance represented by the history table 163, in accordance with the selected policy selection method, if past performance that matches the cluster characteristics of the cluster is included in the history. For example, if multiple policies are identified for the cluster from the policy matching table 502, the policy that is judged to have a relatively high execution effect based on past performance that matches the cluster characteristics may be selected from among those multiple policies. For clusters in which the selected policy has been implemented, the processing execution unit 140 may add past performance for the implemented policy to the history represented by the history table 163. The history represented by the history table 163 may consist of one or more past performances, and each of the one or more past performances may include the implemented policy, the effect of the implementation of the policy, and the cluster characteristics of the cluster in which the target persons to whom the policy was implemented were classified. This is expected to result in the selected policy being one with a higher expected value for achieving the objective.

[0078] Examples of the technical significance of applying environmental / group characteristics to the data processing of the support system 100 are as follows:

[0079] Without applying environmental / group characteristics, the data used for selecting strategies would be similar to that used in typical digital marketing or log analysis (for example, data on how many people accessed a website as a result of a strategy such as creating a website to advertise product X or sending out emails promoting that website). This would simply involve looking at past trends to determine whether or not people are likely to be interested in product X.

[0080] When environmental / group characteristics are applied, aspects that cannot be determined solely from the measures and their effects—specifically, the circumstances in which the target individual finds themselves within the environment / group to which they belong—can be applied to data processing as environmental / group characteristics, thereby enabling the selection of more accurate and effective measures.

[0081] As mentioned above, applying environmental / group characteristics can be expected to improve clustering accuracy. This is because the overall points used for sampling are calculated not only based on the individual's results (individual behavior data) but also on the characteristics of the environment / group in which the individual is placed, expressed as numerical values ​​(e.g., coefficients). Specifically, for example, when clustering purchasing behavior, if only individual behavior data is used, it is expected that clustering will be based only on whether or not the individual responded to the measure or the difference in the number of accesses to the web page. However, by using individual characteristics data (data representing the characteristics of the environment / group to which the individual belongs) integrated with such individual behavior data, it becomes possible to classify the individual into a more appropriate cluster. For example, an individual who would be classified into a high cluster based on behavior data alone may be classified into a lower cluster when individual characteristics data is taken into account (e.g., "want to buy but can't" (e.g., many web page views but no purchase due to organizational problems)).

[0082] Furthermore, the application of environmental / group characteristics is expected to contribute to improving clustering accuracy, as well as to all of the following: The environmental / group characteristics may be automatically classified into either "characteristics suitable for overall transition" or "characteristics suitable for individual transition" either manually or using a model that takes the characteristics of the environmental / group characteristics as input and outputs classification results. Information indicating which characteristics the characteristics were classified into may be included in the corresponding influence 313 or 323 for the environmental / group characteristics. "Characteristics suitable for overall transition" are characteristics that are expected to change over a relatively long period of time (for example, a good work environment), while "characteristics suitable for individual transition" are characteristics that are expected to change in a relatively short period of time due to changeable temporary states or emotional changes (for example, an organization that is easily influenced by its surroundings). The correlation method may be a method that emphasizes "characteristics suitable for overall transition," and the before-and-after method may be a method that emphasizes "characteristics suitable for individual transition." In this way, a policy selection method suitable for the environmental / group characteristics is selected, and it is expected that policies with higher expected effects will be selected. • By applying environmental / group characteristics, the accuracy of clustering and the accuracy of selecting policy selection methods are expected to improve, thereby improving the appropriateness of the selected policies. Specifically, for example, since policies effective for environmental / group characteristics can be estimated in addition to past performance, an improvement in policy accuracy is expected.

[0083] Although one embodiment has been described above, this is merely an example for the purpose of explaining the present invention, and is not intended to limit the scope of the present invention to this embodiment alone. The present invention can be implemented in various other forms.

[0084] For example, in this embodiment, purchasing behavior is used as an example of a cluster behavior indicator, but other behavior indicators may be used, such as application behavior that contributes to the hiring rate or organizational contribution behavior that contributes to organizational productivity.

[0085] For example, instead of or in addition to the overall points, cluster index values ​​may be provided for each of one or more cluster items, and each of several different cluster conditions relating to one or more combinations of cluster index values ​​may be associated with one of several clusters. The clusters are not limited to sequential relationships; at least some of them may be parallel relationships. For example, the clusters may be multiple nodes in a Directed Acyclic Graph (DAG), and the edges in the DAG may represent the relationships between clusters. The cluster corresponding to the starting node in the DAG may be the cluster furthest from the goal, and the cluster corresponding to the ending node in the DAG may be the cluster closest to the goal (or the target cluster). For each subject, cluster index values ​​may be calculated for each of one or more cluster items using the subject's behavioral data and subject's characteristic data, and the subject may be classified into a cluster associated with a cluster condition that fits the calculated combination of one or more cluster index values.

[0086] Furthermore, the multiple clusters 50 may be multiple stages that are generated by the method disclosed in Patent Document 1 and are arranged in stages toward the goal.

[0087] Furthermore, at least a part of the disclosure in Patent Document 1 may be incorporated into (e.g., applied to or adapted to) at least a part of this embodiment (e.g., the before-and-after method). [Explanation of symbols]

[0088] 100... Support System

Claims

1. A data integration unit integrates, for each of multiple subjects, subject behavior data, which is data about the subject from behavioral data that includes various data measured about the behavior of the multiple subjects, and subject characteristic data, which is data about the characteristics of the environment / group to which the subject belongs from characteristic data that is data that follows the characteristics of the environment / group, which is the environment and / or group, and uses the subject behavior data and the subject characteristic data to calculate a cluster index value, which is an index value used for clustering. For each of the aforementioned multiple subjects, a clustering unit classifies the subject into the cluster to which the calculated cluster index value belongs, among multiple clusters moving toward the target behavior, where each subject is associated with an index value range, which is a range of cluster index values. For each of the aforementioned clusters, a policy selection unit selects a policy selection method from among several policy selection methods that is suitable for the cluster characteristics, which are the characteristics of the environment / group to which the target persons classified into that cluster belong, and selects policies for the target persons classified into that cluster according to the selected policy selection method. For each of the aforementioned clusters, a processing execution unit executes result processing, which is processing related to the selected measures. A system equipped with features to support behavioral change in target individuals.

2. For each of the aforementioned multiple subjects, The aforementioned subject behavior data includes quantitative data and qualitative data. The cluster index value for the subject is calculated based on the quantitative data representing each of one or more items of the subject, the value set based on the qualitative data of the subject, and the value set based on the characteristics of the environment / group to which the subject belongs. The system according to claim 1.

3. The aforementioned multiple policy selection methods include correlation methods and precede-precedent methods. The correlation method is a method for selecting measures that have a relatively high correlation with the behaviors represented by the behavioral data of individuals classified into a cluster, for the purpose corresponding to the cluster to which the correlation method was selected. The aforementioned before-and-after method is a method for selecting measures for an objective that corresponds to the gap between the characteristics of the cluster from which the before-and-after method was selected and the characteristics of the cluster with the next highest index value range. For each of the aforementioned clusters, the policy selection unit shall: If the cluster characteristics of the cluster in question are defined as being suitable for overall transition, the correlation method is selected. If the cluster characteristics of the cluster in question are defined as being suitable for individual transitions, then the above-preceding method is selected. The system according to claim 1.

4. If, for at least one of the aforementioned clusters, past performance that matches the cluster characteristics of that cluster is included in the history represented by the historical data, the policy selection unit selects a policy for that cluster based on that past performance, in accordance with the selected policy selection method. Of the aforementioned clusters, for the cluster where the selected measure was implemented, the processing execution unit adds past performance to the history represented by the history data. The history represented by the aforementioned historical data consists of one or more past performance records. Each of the one or more past performances mentioned above includes the measures taken, the effects of the measures taken, and the cluster characteristics of the clusters to which the target persons who took the measures were classified. The system according to claim 1.

5. For each of the multiple subjects, subject behavior data, which is data about that subject from behavioral data that includes various data measured about the behavior of the multiple subjects, and subject characteristic data, which is data about the characteristics of the environment / group to which the subject belongs from characteristic data that is data that follows the characteristics of the environment / group, which is the environment and / or group, are integrated. Using the subject behavior data and subject characteristics data, a cluster index value, which is an index value used for clustering, is calculated. For each of the aforementioned multiple subjects, an index value range, which is a range of cluster index values, is associated with each subject. Among the multiple clusters moving toward the target behavior, the subject is classified into the cluster to which the index value range to which the calculated cluster index value belongs is associated. For each of the aforementioned clusters, a policy selection method is selected from among several policy selection methods that is suitable for the cluster characteristics, which are the characteristics of the environment / group to which the target persons classified into that cluster belong. In accordance with the selected policy selection method, policies will be selected for the target individuals classified into the cluster. For each of the aforementioned clusters, a result processing operation is performed, which is the processing related to the selected measures. A method of supporting behavioral change in individuals by having a computer perform certain actions.

Images