Causality exploration device

Abstract

To reduce the number of attributes of customers subjected to causal search processing while maintaining the accuracy of the causal search processing, thereby outputting a graph in which a causal relation between the attributes can be discriminated.SOLUTION: A causal search device 10 comprises: a layering unit 12 that extracts a B group consisting of customers who have completed procedures on the WEB and an A group consisting of customers who have visited the WEB but have completed procedures in a place other than the WEB, and divides the A group into a plurality of layers along a line of flow on the WEB; an attribute extraction unit 13 that, based on the feature amount for every attribute of the layers obtained from information on the attributes of the customers in the layers of the A group and the customers in a CV layer forming the B group, extracts, for each of the layers of the A group, a correlation attribute not having correlation with the attribute of the customers in the CV layer but having correlation with the attribute of the customers in any one of the layers of the A group; a causal search unit 14 that executes the causal search between the correlation attributes for each of the layers of the A group; and a graph generation and output unit 15 that generates and outputs a directed acyclic graph representing a result of the causal search for each of the layers of the A group.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a causal exploration device that performs causal exploration between attributes of target customers. Note that causal exploration between attributes means exploring (estimating) the causal relationship between attributes, and the causal relationship between attributes means a relationship in which one of two attributes is the cause and the other is the result.

Background Art

[0002] In recent years, customers have been able to execute the same contract procedures on a web page (hereinafter also referred to as a "WEB page") as those performed in a store. However, even if a customer who needs to perform a contract procedure visits a WEB page, they do not always execute the contract procedure on the WEB page, and may leave the WEB page and perform the contract procedure at an actual store. In order to find out the factors for a customer to leave the WEB page, it is effective to extract the attributes specific to customers who left the WEB page and performed the contract procedure at an actual store, and analyze the causal relationship between those specific attributes. When analyzing such a causal relationship, a technique called causal exploration (or cause exploration) is used (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the above causal relationship exploration process, when the number of customer attributes to be processed increases, the number of nodes and edges of the graph (generally a Directed Acyclic Graph (DAG)) representing the causal relationships between attributes becomes extremely large, resulting in the graph becoming huge. As a result, inconveniences such as difficulty in determining the causal relationships between attributes from the graph may occur. On the other hand, simply reducing the number of customer attributes to be processed in the causal relationship exploration process raises concerns about a decrease in the accuracy of the causal relationship exploration process.

[0005] The present disclosure has been made to solve the above problems, and aims to output a graph capable of determining causal relationships between attributes by reducing the number of customer attributes to be processed in the causal relationship exploration process while maintaining the accuracy of the causal relationship exploration process.

Means for Solving the Problems

[0006] The causal relationship exploration device according to the present disclosure extracts, from a plurality of customers who have completed a certain procedure, a first customer group that visited the specific procedure location and completed the procedure at the specific procedure location and a second customer group that visited the specific procedure location but completed the procedure at a location other than the specific procedure location, based on information regarding the procedure completion location for each customer and information regarding whether or not they visited a specific procedure location. Then, based on the visit history information for each customer, a stratification unit stratifies the second customer group into a plurality of layers defined along the route until the procedure is completed at the specific procedure location. The device acquires the attribute information of each customer in each layer of the second customer group stratified by the stratification unit and the customers in the conversion layer that constitutes the first customer group extracted by the stratification unit. Based on the feature amounts for each attribute of each layer of the second customer group and the conversion layer obtained based on the acquired attribute information, an attribute extraction unit extracts a correlation attribute that has no correlation with the attributes of the customers in the conversion layer but has a correlation with the attributes of the customers in any layer of the second customer group, for each layer of the second customer group. Based on the correlation attributes for each layer of the second customer group extracted by the attribute extraction unit and the feature amounts for each attribute of each layer of the second customer group and the conversion layer, a causal relationship exploration unit performs a causal relationship exploration among the correlation attributes for each layer of the second customer group. A graph generation and output unit generates and outputs a graph representing the causal relationship exploration result for each layer of the second customer group obtained by the causal relationship exploration by the causal relationship exploration unit.

[0007] Note that the above-mentioned "conversion layer" means, for example, a layer composed of customers who have been converted from "prospective customers" to "customers" by completing a procedure at a specific procedure location.

[0008] In the above causal exploration device, the stratification unit extracts, from a plurality of customers who have completed a certain procedure, a first customer group that visited a specific procedure location and completed the procedure at the specific procedure location and a second customer group that visited the specific procedure location but completed the procedure outside the specific procedure location, based on information regarding the location where each customer completed the procedure and information on whether or not they visited a specific procedure location. Then, based on the visit history information for each customer, the second customer group is stratified into a plurality of layers defined along the flow line until the procedure is completed at the specific procedure location. As a result, the second customer group is stratified (grouped) into a plurality of layers defined along the flow line until the procedure is completed at the specific procedure location. Next, the attribute extraction unit acquires the respective attribute information of the customers in each layer of the second customer group and the customers in the conversion layer that constitutes the first customer group, and based on the feature amounts for each attribute of each layer of the second customer group and the conversion layer obtained based on the acquired attribute information, extracts, for each layer of the second customer group, a correlation attribute that has no correlation with the attributes of the customers in the conversion layer but has a correlation with the attributes of the customers in any layer of the second customer group. As a result, based on the attribute information of each layer of the second customer group and the conversion layer respectively, correlation attributes (attributes that have no correlation with the attributes of the customers in the conversion layer but have a correlation with the attributes of the customers in any layer of the second customer group) that are the targets of subsequent "causal exploration" are appropriately extracted for each layer of the second customer group. Then, the causal exploration unit executes causal exploration between the correlation attributes for each layer of the second customer group extracted, and based on the feature amounts for each attribute of each layer of the second customer group and the conversion layer respectively. Furthermore, the graph generation output unit generates and outputs a graph representing the causal exploration results for each layer of the second customer group obtained by the causal exploration. As described above, since causal exploration is executed for each layer of the second customer group with the appropriately extracted correlation attributes as the targets, the accuracy of the causal exploration process can be maintained at a certain level or higher, and by reducing the number of customer attributes that are the targets of a single causal exploration process, a graph capable of discriminating the causal relationship between attributes can be output.

Advantages of the Invention

[0009] According to the present disclosure, it is possible to output a graph capable of discriminating causal relationships between attributes by reducing the number of customer attributes to be processed in the causal search process while maintaining the accuracy of the causal search process.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Embodiments for Carrying Out the Invention

[0011] With reference to the accompanying drawings, an embodiment of the causal exploration device will be described. Hereinafter, as an example of the "procedure", the contract procedure of a certain target service will be described, and it is assumed that the contract procedure by the customer can be performed either on the WEB or at an actual store as the place where the procedure is completed.

[0012] As shown in FIG. 1, a causal exploration device 10 according to an embodiment includes a customer data holding unit 11, a stratification unit 12, an attribute extraction unit 13, a causal exploration unit 14, and a graph generation / output unit 15. Hereinafter, the functions of each unit will be described.

[0013] The customer data holding unit 11 is a database that holds information about the completion location of the procedure for each customer, information on whether or not to access a group of WEB pages for the procedure, the accessed page information among the group of WEB pages for the procedure, the attribute information of the customer, etc., as described below. The above various types of information are obtained in advance from the terminal logs, application logs, subscriber information, etc. of the customers who have given consent to obtain them, and are held by the customer data holding unit 11.

[0014] The holding forms of the above various types of information can take various forms, and an example will be described using FIGS. 2(a) to 2(c). As shown in FIG. 2(a), information regarding the location where the procedure is completed for each customer and information regarding the presence or absence of access to the group of WEB pages for the procedure are held by the customer data holding unit 11 using the customer identification information (customer ID) as a key, together with information on the event occurrence time (here, the time when the contract procedure is completed). The above-mentioned "information regarding the location where the procedure is completed" is recorded as "WEB" when the customer completes the contract procedure on the WEB, for example, and is recorded as "REAL" when the customer completes the contract procedure at an actual store. The above-mentioned "information regarding the presence or absence of access to the group of WEB pages for the procedure" is recorded as "1" when there is access, for example, and is recorded as "0" when there is no access. The "group of WEB pages for the procedure" here includes various WEB pages that are assumed to be accessed in a series of processes when the customer conducts the contract procedure for the target service. For example, it includes pages related to the fee plan of the target service, pages related to the comparison of fee plans, the contract procedure start page, the page during the contract procedure, the contract procedure completion page, pages related to FAQ (Frequently Asked Questions), and the like. The information shown in FIG. 2(a) is collectively referred to as procedure-related information 11A. As shown in FIG. 2(b), the access page information for each customer is held by the customer data holding unit 11 using the customer ID as a key, together with information on the event occurrence time (here, the time when the corresponding WEB page is accessed). The information shown in FIG. 2(b) is collectively referred to as access-related information 11B. The attribute information 11C for each customer shown in FIG. 2(c) is held by the customer data holding unit 11 using the customer ID as a key, with the content of the attribute or a flag indicating whether the customer has the attribute in columns for each of various attributes. For example, for attributes A and B, a flag "1" indicating that the customer has the attribute or a flag "0" indicating that the customer does not have the attribute is held, and for attribute C, the content of the attribute C such as YYY, ZZZ, XXX is held.Note that the information on "event occurrence time" shown in FIGS. 2(a) and 2(b) is used to determine whether the corresponding contract procedure completion and web page access were performed as a series of operations. For example, if the time of web page access and the time of contract procedure completion are separated by a predetermined time interval or more, it is determined that they were not performed as a series of operations.

[0015] Returning to FIG. 1, the stratification unit 12 extracts, from a plurality of customers who have completed the contract procedure for a certain target service, a first customer group (hereinafter referred to as "Group B") consisting of customers who have visited the web page group for procedures and completed the procedures on the web page group for procedures, and a second customer group (hereinafter referred to as "Group A") consisting of customers who have visited the web page group for procedures but completed the procedures outside the web page group for procedures. Based on the access history information for each customer, it is a functional unit that stratifies the customers belonging to Group A (hereinafter abbreviated as "Group A") into a plurality of layers defined along the flow line until the procedures are completed in the web page group for procedures. Although details will be described later, the stratification unit 12 obtains the flow line until the customers belonging to Group B (hereinafter abbreviated as "Group B") complete the procedures on the web page group for procedures, and based on both the obtained flow line of Group B and a predetermined reference flow line serving as a reference, determines a plurality of layers (hereinafter also referred to as "stratification variables") for stratifying Group A, and uses these plurality of layers for stratifying Group A. However, it is not essential to use the flow line of Group B as described above, and the stratification variables may be determined based only on the predetermined reference flow line. Also, the process of determining the stratification variables may be omitted, and a plurality of layers corresponding to the predetermined reference flow line may be used for stratifying Group A.

[0016] The attribute extraction unit 13 acquires the attribute information of each customer in each layer of group A stratified by the stratification unit 12 and the customers in the conversion layer (hereinafter also referred to as "layer CV"), which is the layer constituting group B extracted by the stratification unit 12. Based on the feature amounts for each attribute of each layer of group A and the conversion layer obtained based on the acquired attribute information, the correlation attribute is an attribute that has no correlation with the attributes of the customers in the conversion layer but has a correlation with the attributes of the customers in any layer of group A. The attribute extraction unit 13 is a functional unit that extracts the correlation attributes for each layer of group A. Although details will be described later, the attribute extraction unit 13 tabulates the number of customers having the attribute related to the attribute information for the attribute information of each layer of group A and the conversion layer, and acquires the obtained number of customers (total value) as the feature amount for each attribute. At this time, in addition to using the number of customers obtained by tabulation as the feature amount as it is, the calculation result of performing a predetermined calculation such as multiplying the number of customers by a predetermined coefficient may be acquired as the feature amount. Further, the attribute extraction unit 13 excludes attributes for which the number of customers having a certain attribute is below a predetermined lower limit standard from the correlation attributes. Such a processing example related to attribute extraction will be described in detail later.

[0017] The causal exploration unit 14 is a functional unit that performs causal exploration between the correlation attributes for each layer of group A extracted by the attribute extraction unit 13, based on the feature amounts for each attribute of each layer of group A and the conversion layer.

[0018] The graph generation and output unit 15 is a functional unit that generates and outputs a directed acyclic graph (DAG) representing the causal exploration result for each layer obtained by the causal exploration by the causal exploration unit 14.

[0019] Next, the processing executed in the causal exploration device 10 will be sequentially described along the flowchart of FIG. 3. The processing in FIG. 3 may be started, for example, at a predetermined periodic timing, or may be started triggered by a predetermined start operation or the like performed by the operator of the causal exploration device 10.

[0020] The stratification unit 12 obtains, by reading out the procedure-related information 11A in Fig. 2(a) regarding a plurality of customers (hereinafter referred to as "customers with completed procedures") who have completed the contract procedures for the target service from the customer data storage unit 11 (step S1), and extracts group A and group B from the customers with completed procedures as follows based on the procedure-related information 11A (step S2). That is, as shown in Fig. 5, the stratification unit 12 extracts, as group A, a customer (customer ID: 0000001) whose procedure completion location in the procedure-related information 11A is "REAL (an actual store other than the WEB)" and whose access presence / absence information to the procedure-related WEB page group is "1 (accessed)", and extracts, as group B, a customer (customer ID: 0000010) whose procedure completion location is "WEB" and whose access presence / absence information to the procedure-related WEB page group is "1 (accessed)". Note that customers other than the above, that is, a customer (customer ID: 0000002) whose procedure completion location is "REAL" and whose access presence / absence information to the procedure-related WEB page group is "0 (not accessed)" is excluded from the extraction target.

[0021] Next, the stratification unit 12 obtains the access-related information 11B for each of group A and group B shown in Fig. 6 by reading out the access-related information 11B in Fig. 2(b) from the customer data storage unit 11 for each of the extracted group A and group B (step S3).

[0022] Next, the stratification unit 12 determines whether the stratification variable to be used in the stratification in step S6 described later has been determined (step S4). Here, for example, when continuously and repeatedly executing the same causal search process (Figure 3) for the target service, if the stratification variable determined in the first process can be reused in the second and subsequent processes, it is determined that the stratification variable to be used has been determined. In other cases, it is determined that the stratification variable to be used has not been determined. If it is determined in step S4 that the stratification variable has been determined, the process proceeds to step S6 described later. On the other hand, if it is determined that the stratification variable has not been determined, the stratification variable is determined as follows (step S5). As shown in Figure 7, the WEB access history information of group B is applied to the predefined main flow (framework). Here, as the main flow, frameworks such as layer A, layer B, layer C, layer D, and layer CV are predefined. Layer A is the layer of pages related to the fee plan, layer B is the layer of pages for comparing fee plans, layer C is the layer of the WEB procedure start page, layer D is the layer of pages during the WEB procedure, and layer CV is the layer of the WEB procedure completion page. Each layer is determined for each category of WEB pages. Pages other than the above (for example, pages related to FAQ) are not included in the main flow (framework) and are not considered in the following stratification variable determination process. On the left side of Figure 7, the access history information of customers with customer IDs "0000010" and "0000011" belonging to group B is illustrated. In the example of the accessed pages, "aa, aaa" are pages related to the fee plan, "bbb" is a page for comparing fee plans, "ccc" is the WEB procedure start page, "dd, ddd" are pages during the WEB procedure, and "cvpage" is the WEB procedure completion page. At this time, as shown in Figure 7, the flow of the customer with customer ID "0000010" until the contract is completed is found to be "layer A → layer D → layer CV", and the flow of the customer with customer ID "0000011" until the contract is completed is found to be "layer A → layer B → layer C → layer D → layer CV". The flow "layer A → layer B → layer C → layer D → layer CV" that comprehensively includes the flows of customers belonging to group B as described above is derived as the flow until group B completes the contract, and five variables, "layer A, layer B, layer C, layer D, layer CV", are determined as the stratification variables.

[0023] Next, the stratification unit 12 stratifies Group A as follows using the stratification variable determined in step S5 or already determined (step S6). At this time, when visiting across multiple layers, according to the rule of adopting the layer closer to layer CV among the multiple layers, as shown in FIG. 8, the customer with customer ID "0000001" visits only the pages assigned to layer A, so is stratified into layer A, and the customer with customer ID "0000003" visits the page assigned to layer B as the page closest to layer CV, so is stratified into layer B. Similarly, the customer with customer ID "0000004" is stratified into layer C, and the customer with customer ID "0000005" is stratified into layer D.

[0024] Next, the stratification unit 12 generates a layer information table shown on the right side of FIG. 9 that records the stratification result (information on the stratified layer) for each customer in association with the customer ID (step S7), and passes it to the attribute extraction unit 13.

[0025] Next, the attribute extraction unit 13 obtains the customer attribute information 11C in FIG. 2(c) by reading it from the customer data holding unit 11 (step S8), and generates an attribute table with layer information shown in FIG. 10 by matching the layer information table (FIG. 9) received from the stratification unit 12 with the customer attribute information (step S9). As shown in FIG. 10, in the attribute table with layer information, information regarding the layer to which the customer belongs and various attributes that the customer has is held with the customer ID as the key.

[0026] Next, the attribute extraction unit 13 executes an extraction process of extracting a "correlation attribute", which has no correlation with the attributes of the customers in layer CV but has a correlation with the attributes of the customers in any of the layers in Group A, for each layer in Group A (step S10). The extraction process in step S10 is represented by the subroutine in FIG. 4. First, the attribute extraction unit 13 totals the total number of people belonging to each layer and the number of people having each attribute for each of layers A to D and layer CV, and uses the number of people having each attribute (total value) as the feature amount of the attribute in that layer (step S10A). Thereby, as shown in FIG. 11, the total number of people belonging to each layer and the number of people having each attribute (feature amount of the attribute) for each layer are obtained.

[0027] Next, the attribute extraction unit 13 calculates the ratio of the number of people with each attribute for each layer by substituting numerical values into the formula shown in the upper right of FIG. 12 (step S10B). For example, for attribute A in layer A, the ratio of the number of people is (9000 / 13800), and 65.2% is obtained. In this way, for each attribute for each layer, the ratio of the number of people as shown in FIG. 12 is obtained.

[0028] Next, the attribute extraction unit 13 compares the ratio of the number of people with each attribute in layers A to D with the ratio of the number of people with each attribute in layer CV, and extracts, as correlation attributes, the attributes for which the ratio of the number of people is twice or more the ratio of the number of people in layer CV among each attribute in layers A to D (step S10C). For example, as shown in FIG. 13, for "attribute A: 1", since the ratio of the number of people in layer A, 34.8%, and the ratio of the number of people in layer B, 27.7%, are both twice or more the ratio of the number of people in layer CV, 13.2%, "attribute A: 1" is extracted as a correlation attribute correlated with layers A and B. Similarly, for "attribute B: 0", since the ratio of the number of people in each of layers A to D is twice or more the ratio of the number of people in layer CV, 1.3%, "attribute B: 0" is extracted as a correlation attribute correlated with layers A to D. Here, as the extraction condition for correlation attributes, the condition "the ratio of the number of people is twice or more the ratio of the number of people in layer CV" is used, but this condition is just an example, and another condition such as "the ratio of the number of people is five times or more the ratio of the number of people in layer CV" may also be used.

[0029] Furthermore, the attribute extraction unit 13 excludes, from the correlation attributes extracted in step S10C, attributes for which the number of people having the attribute is below a predetermined lower limit criterion (step S10D). Here, as an example, the attribute extraction unit 13 excludes, among the extracted correlation attributes, attributes for which the number of people having the attribute is 1% or less of the total number of people in the layer. For the layer D shown in FIG. 14, "Attribute B: 0" is excluded from the correlation attributes because the number of people having the attribute, 61, is 1% or less of the total number of people in layer D, 1149. Note that the condition here, "the number of people having the attribute is 1% or less of the total number of people in the layer", is an example, and another condition such as "the number of people having the attribute is 3% or less of the total number of people in the layer" may be used.

[0030] Through the processing of FIG. 4 as described above, among the attributes for each of layers A to D, the attributes shown in FIG. 15 are extracted as correlation attributes (attributes that have no correlation with layer CV but have a correlation with the layer and have a sufficient number of people having the attribute). Note that FIG. 15 also shows correlation attributes related to attributes C to G not illustrated in FIGS. 11 to 14.

[0031] Next, the causal relationship exploration unit 14 performs causal relationship exploration for each layer (step S11). Using FIG. 16, for example, to explain the causal relationship exploration for layer A, the causal relationship exploration unit 14 performs causal relationship exploration on a total of four attributes, namely, " Attribute A: 1", " Attribute B: 0", " Attribute C: XXX", with ": Layer A" added. As a result, out of a total of 12 possible causal relationships between attributes shown in the lower center of FIG. 16 as candidates, a total of 6 causal relationships shown on the lower right side of FIG. 16 are obtained as a result of the causal relationship exploration.

[0032] Next, the graph generation and output unit 15 generates (step S12) and outputs (step S13) a directed acyclic graph (DAG) representing the causal relationship exploration result for each layer obtained through the causal relationship exploration. For example, for layer A, as shown in FIG. 17, a directed acyclic graph representing a total of 6 causal relationships obtained through the causal relationship exploration is generated and output.

[0033] According to one embodiment of the causal relationship exploration device 10 described above, causal relationship exploration is performed for each layer of group A based on the correlation attributes appropriately extracted for each layer of group A and each feature for each attribute of layer CV. Therefore, the accuracy of the causal relationship exploration process can be maintained at a certain level or higher, and the load of the causal relationship exploration process can be reduced by reducing the number of customer attributes to be processed in one causal relationship exploration process.

[0034] In addition, since causal relationship exploration is performed for each layer of group A and the number of customer attributes to be processed in one causal relationship exploration process can be reduced, it is possible to avoid the directed acyclic graph representing the causal relationship between attributes from becoming huge, and a graph capable of discriminating the causal relationship between attributes can be output.

[0035] In the above embodiment, as an example of the "procedure", an example of the contract procedure for a certain target service that can be executed either on the WEB or at an actual store is described. In this case, the results of causal relationship exploration can be effectively used for the analysis of the dropout locations and customer attributes in the flow line until the procedure is completed for group A (the group of customers who visited the group of WEB pages for procedures but completed the procedure outside the group of WEB pages for procedures).

[0036] In addition, the stratification unit 12 determines a plurality of layers (stratification variables) for stratifying group A based on a reference flow line serving as a predetermined reference, thereby preventing, for example, the visit history to a "page of reference information not directly related to the procedure" such as an FAQ page from being used as a basis in advance, and enabling appropriate setting of stratification variables based on the predetermined reference flow line.

[0037] In addition to the predetermined reference flow line, the stratification unit 12 further determines a plurality of layers (stratification variables) for stratifying group A based on the flow line of group B until the procedure is completed on the group of WEB pages for procedures, so that more appropriate stratification variables closer to the actual situation can be determined by using the flow line of group B that actually completed the procedure on the group of WEB pages for procedures.

[0038] In addition, for the attribute information of each layer in Group A and the CV layer, the attribute extraction unit 13 aggregates the number of customers having the attribute related to the attribute information, and based on the obtained number of customers, acquires the feature amount of each attribute. Thus, based on the number of customers having the attribute (aggregated value), the feature amount of each attribute can be acquired by relatively simple processing.

[0039] In addition, by the attribute extraction unit 13 excluding an attribute whose number of customers having a certain attribute is below a predetermined lower limit criterion from the correlation attributes, it is possible to prevent a situation where an attribute with an extremely small number of customers having the attribute is extracted as a correlation attribute, and it is possible to optimize the extraction process of the correlation attributes.

[0040] In addition, in the above-described embodiment, an example has been described in which the causal exploration device 10 includes the customer data holding unit 11 that holds information regarding the place where the procedure is completed, information on whether or not to visit a specific procedure location, visit history information, and attribute information for each customer. In this way, since the causal exploration device 10 has the customer data holding unit 11 inside, it becomes unnecessary to acquire various necessary information from the outside when executing the process of FIG. 3, and it is possible to contribute to speeding up the process.

[0041] (Explanation of terms, explanation of hardware configuration (FIG. 18), etc.) Note that the block diagrams used in the description of the above embodiment and modification examples show blocks of functional units. These functional blocks (components) are realized by an arbitrary combination of at least one of hardware and software. Also, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one physically or logically combined device, or two or more physically or logically separated devices may be directly or indirectly (for example, using wired, wireless, etc.) connected and realized using these multiple devices. The functional block may be realized by combining software with the above one device or the above multiple devices.

[0042] Functions include, but are not limited to, judgment, decision-making, determination, calculation, computation, processing, derivation, investigation, search, confirmation, reception, transmission, output, access, solution, selection, selection determination, establishment, comparison, assumption, expectation, regarded as, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), assigning, etc. For example, a functional block (component) that enables transmission is called a transmitting unit or a transmitter. As described above, the implementation method is not particularly limited.

[0043] For example, the causal search device in an embodiment of the present disclosure may function as a computer that performs the processing in this embodiment. FIG. 18 is a diagram showing a hardware configuration example of a causal search device 10 according to an embodiment of the present disclosure. Physically, the above-described causal search device 10 may be configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, and the like.

[0044] In the following description, the term "device" can be read as a circuit, a device, a unit, etc. The hardware configuration of the causal search device 10 may be configured to include one or more of each device shown in the figure, or may be configured without including some devices.

[0045] Each function in the causal search device 10 is realized by causing the processor 1001 to perform operations by loading a predetermined software (program) onto hardware such as the processor 1001 and the memory 1002, controlling communication by the communication device 1004, and controlling at least one of reading and writing data in the memory 1002 and the storage 1003.

[0046] Processor 1001 controls the entire computer by operating, for example, an operating system. Processor 1001 may be composed of a central processing unit (CPU: Central Processing Unit) including an interface with peripheral devices, a control device, an arithmetic device, registers, etc.

[0047] Also, processor 1001 reads a program (program code), software module, data, etc. from at least one of storage 1003 and communication device 1004 into memory 1002, and executes various processes according to these. As the program, a program that causes a computer to execute at least a part of the operations described in the above embodiments is used. Although it has been described that the above various processes are executed by one processor 1001, they may be executed simultaneously or sequentially by two or more processors 1001. Processor 1001 may be implemented by one or more chips. Note that the program may be transmitted from a network via a telecommunication line.

[0048] Memory 1002 is a computer-readable recording medium and may be composed of, for example, at least one of ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 1002 may be referred to as a register, cache, main memory (main storage device), etc. Memory 1002 can store a program (program code), software module, etc. executable for implementing the wireless communication method according to an embodiment of the present disclosure.

[0049] Storage 1003 is a computer-readable recording medium and may be composed of at least one of, for example, an optical disc such as a CD-ROM (Compact Disc ROM), a hard disk drive, a flexible disk, a magneto-optical disk (e.g., a compact disc, a digital versatile disc, a Blu-ray (registered trademark) disc), a smart card, a flash memory (e.g., a card, a stick, a key drive), a floppy (registered trademark) disk, a magnetic strip, etc. Storage 1003 may be referred to as an auxiliary storage device. The above-described storage medium may be, for example, a database including at least one of the memory 1002 and the storage 1003, or other appropriate media.

[0050] The communication device 1004 is hardware (a transmission / reception device) for performing communication between computers via at least one of a wired network and a wireless network, and is also referred to as, for example, a network device, a network controller, a network card, a communication module, etc.

[0051] The input device 1005 is an input device (e.g., a keyboard, a mouse, a microphone, a switch, a button, a sensor, etc.) that receives an external input. The output device 1006 is an output device (e.g., a display, a speaker, an LED lamp, etc.) that performs an output to the outside. Note that the input device 1005 and the output device 1006 may have an integrated configuration (e.g., a touch panel). Also, each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus or may be configured using different buses for each device.

[0052] Each aspect / embodiment described in the present disclosure may be used alone, in combination, or may be switched and used during execution. Also, the notification of predetermined information (e.g., the notification of "being X") is not limited to being explicitly performed, and may be performed implicitly (e.g., by not performing the notification of the predetermined information).

[0053] As described above in detail, it is obvious to those skilled in the art that the present disclosure is not limited to the embodiments described in the present disclosure. The present disclosure can be implemented as modifications and variations without departing from the spirit and scope of the present disclosure as defined by the claims. Therefore, the description of the present disclosure is for illustrative purposes and does not have any limiting meaning for the present disclosure.

[0054] The processing procedures, sequences, flowcharts, etc. of each aspect / embodiment described in the present disclosure may be interchanged as long as there is no contradiction. For example, regarding the methods described in the present disclosure, the elements of various steps are presented using an exemplary order and are not limited to the specific order presented.

[0055] The input / output information, etc. may be stored in a specific location (e.g., memory) or may be managed using a management table. The input / output information, etc. may be overwritten, updated, or appended. The output information, etc. may be deleted. The input information, etc. may be transmitted to other devices.

[0056] In the present disclosure, the description "based on" does not mean "only based on" unless otherwise specified. In other words, the description "based on" means both "only based on" and "at least based on".

[0057] In the present disclosure, when the terms "include", "including" and their variations are used, these terms are intended to be inclusive, similar to the term "comprising". Furthermore, the term "or" used in the present disclosure is not intended to be an exclusive disjunction.

[0058] In the present disclosure, for example, when articles are added by translation, such as a, an, and the in English, the present disclosure may include that the nouns following these articles are in the plural form.

[0059] In the present disclosure, the term "A and B are different" may mean that "A and B are different from each other". Note that the term may also mean that "A and B are each different from C". Terms such as "separate", "coupled", etc. may also be interpreted in the same way as "different".

Explanation of Reference Numerals

[0060] 10... Causality exploration device, 11... Customer data holding unit, 12... Stratification unit, 13... Attribute extraction unit, 14... Causality exploration unit, 15... Graph generation output unit, 1001... Processor, 1002... Memory, 1003... Storage, 1004... Communication device, 1005... Input device, 1006... Output device, 1007... Bus.

Claims

1. From a plurality of customers who have completed a certain procedure, based on information regarding the location where each customer completed the procedure and information on whether or not they visited a specific procedure location, a first customer group that visited the specific procedure location and completed the procedure at the specific procedure location, and a second customer group that visited the specific procedure location but completed the procedure outside the specific procedure location are extracted. Based on the visit history information for each customer, among a plurality of layers defined along the flow path until the procedure is completed at the specific procedure location, which correspond to a plurality of stages on the flow path until the procedure is completed, each customer belonging to the second customer group is classified into the layer corresponding to the stage closest to the completion of the procedure that each customer belonging to the second customer group visited at the specific procedure location, thereby stratifying the second customer group into a stratification unit; Obtaining the respective attribute information of each customer in each layer of the second customer group stratified by the stratification unit and the customers in the conversion layer constituting the first customer group extracted by the stratification unit, aggregating the number of customers having the attribute related to the respective obtained attribute information, and based on the feature amount for each attribute of each layer of the second customer group and the conversion layer obtained as the result of a predetermined calculation performed on the obtained number of customers or the number of customers, an attribute extraction unit that extracts a correlation attribute that has no correlation with the attributes of the customers in the conversion layer but has a correlation with the attributes of the customers in any layer of the second customer group; A causal search unit that executes a causal search for each layer of the second customer group, targeting the correlation attributes for each layer of the second customer group extracted by the attribute extraction unit and the attribute indicating that they belong to each layer, and extracting a relationship that corresponds to a causal relationship among the relationships of two attributes assumed in the plurality of target attributes, where one is the cause and the other is the result; A graph generation output unit that generates and outputs a graph representing the causal search result for each layer of the second customer group obtained by the causal search by the causal search unit; A causal search device comprising the above.

2. The stratification unit is From a plurality of customers who have completed a certain procedure, based on information regarding the location where each customer completed the procedure and information regarding whether or not they visited a group of web pages for the procedure, which are the specific procedure locations, a first customer group that visited the group of web pages for the procedure and completed the procedure on the group of web pages for the procedure, and a second customer group that visited the group of web pages for the procedure but completed the procedure outside the group of web pages for the procedure are extracted. Based on the visit history information for each customer, among a plurality of layers defined along the flow line until the procedure is completed on the group of web pages for the procedure, which correspond to a plurality of stages on the flow line until the procedure is completed, each customer belonging to the second customer group is classified into the layer corresponding to the stage closest to the completion of the procedure that each customer belonging to the second customer group visited on the group of web pages for the procedure, thereby stratifying the second customer group. The causal exploration device according to claim 1.

3. The stratifying unit determines the plurality of layers for stratifying the second customer group based on a reference flow line that is a predetermined reference. The causal exploration device according to claim 2.

4. The stratifying unit further determines the plurality of layers based on the flow line until the customers in the first customer group complete the procedure on the group of web pages for the procedure. The causal exploration device according to claim 3.

5. The attribute extraction unit excludes the attribute for which the number of customers having a certain attribute is below a predetermined lower limit criterion from the correlation attributes. The causal exploration device according to any one of claims 1 to 4.

6. A customer data holding unit that holds information regarding the location where the procedure was completed, information regarding whether or not the customer visited a specific procedure location, visit history information, and attribute information for each customer. The causal exploration device according to any one of claims 1 to 5, further comprising the above.

Images