Analytical device

Abstract

To derive highly useful data for prompting customers to change their actions.SOLUTION: An analyzer 10 comprises: a route setting section 11 that sets, in an environment where a first customer group (for example, an off-line group) including a plurality of customers having a first feature and a second customer group (for example, an on-line group) including a plurality of customers having a second feature different from the first feature are present, a route from a feature point in a first distribution to a feature point in a second distribution, from the first distribution in a two-dimensional coordinate system, which is obtained by two-dimensionally compressing a multidimensional vector representing action data and attribute data of the first customer group, and the second distribution in the two-dimensional coordinate system, which is obtained by two-dimensionally compressing a multidimensional vector representing action data and attribute data of the second customer group; and a derivation section 12 that derives data on a plurality of passing points on the set route as data representing a stepwise change from the first customer group to the second customer group.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to an analysis apparatus that performs data analysis for a shift from a first customer group to a second customer group in an environment where there are a first customer group and a second customer group to which a plurality of customers with different characteristics belong.

Background Art

[0002] When a customer performs procedures such as purchasing a product or changing contract terms, in recent years, online procedures performed through a website and offline procedures performed by visiting an actual store or calling a call center are assumed. Among these, online procedures can reduce fees and waiting times compared to offline procedures, which is advantageous for customers in terms of time and money. For companies, if the number of online procedure customers increases, it is also advantageous in that it can improve the efficiency of manned operations for offline procedures.

[0003] Therefore, assuming that there is a certain correlation between customer behavior and the procedure method used by the customer (referred to as "channel" in this case) (here, online procedure or offline procedure), after analyzing these correlations, it has been considered to encourage the customer to change from an offline procedure to an online procedure. As an example of promoting such a change in human behavior, for example, Non-Patent Document 1 describes a technique of using a radar chart to visually represent the difference between the current state and the preferred state of a certain user regarding a plurality of elements for the support of health management, and prompting the user to reach the preferred state.

Prior Art Documents

Non-Patent Documents

[0004] <着

Non-Patent Document 1

Summary of the Invention

[0005] However, while Non-Patent Document 1 allows for the identification of elements with significant differences between the current state and the desired state (elements that need improvement), it is difficult to obtain insights into which elements among the multiple elements represented in the radar chart should be changed and to what extent. Therefore, there is room for improvement in terms of its usefulness in encouraging behavioral change in users. Applying this to the transition from offline procedures to online procedures, it is equivalent to not being able to obtain insights into which behaviors or attributes, such as the types of web pages customers should view or the services they subscribe to, should be changed. Similarly, there is room for improvement in terms of its usefulness in encouraging behavioral change in customers.

[0006] This disclosure is made to address the above-mentioned issues and aims to derive data that is highly useful in encouraging behavioral change in customers. [Means for solving the problem]

[0007] The analytical device relating to this disclosure comprises, in an environment in which a first customer group including a plurality of customers having a first characteristic and a second customer group including a plurality of customers having a second characteristic different from the first characteristic exist, a path setting unit that sets a path from a feature point in the first distribution to a feature point in the second distribution from a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the same two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions, and a derivation unit that derives data relating to a plurality of passing points on the path set by the path setting unit as data representing a stepwise change from the first customer group to the second customer group.

[0008] In the above-described analysis device, in an environment where a first customer group and a second customer group exist, the route setting unit sets a route from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral and attribute data of the first customer group into two dimensions, and a second distribution in the same two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral and attribute data of the second customer group into two dimensions. The derivation unit then derives data related to multiple passing points on the route set by the route setting unit as data representing the gradual change from the first customer group to the second customer group. In this way, data on multiple points along the path from feature points in the first distribution to feature points in the second distribution is derived as data representing the gradual change from the first customer group to the second customer group. Therefore, it becomes possible to obtain insights into which behavioral data and attribute data should be changed and to what extent when customers transform from the first customer group to the second customer group, and it is possible to derive highly useful data for encouraging customers to change their behavior from the first customer group to the second customer group. [Effects of the Invention]

[0009] According to this disclosure, it is possible to derive data that is highly useful in encouraging behavioral change in customers. [Brief explanation of the drawing]

[0010] [Figure 1] This is a functional block diagram of an analytical apparatus according to an embodiment of the invention. [Figure 2] This is a flowchart showing the processes performed in the analytical instrument. [Figure 3] This figure shows an example of data related to customer behavior. [Figure 4] This figure shows a matrix created by transforming the data in Figure 3 to represent the time-series behavior of each user. [Figure 5] This figure shows an example of converting the elements of the matrix in Figure 4 into numerical values. [Figure 6] This figure shows an example of data regarding customer attributes. [Figure 7] A diagram showing an example of converting non-numerical variables in the data of FIG. 6 into numerical values. [Figure 8] A diagram showing which channel, online procedure or offline procedure, a customer used to perform the procedure. [Figure 9] A diagram showing an example of converting offline to 0 and online to 1 into numerical values in the data of FIG. 8. [Figure 10] A diagram representing a deep learning network used in an analyzer. [Figure 11] A diagram showing an example of plotting N customer data on a two-dimensional plane with the outputs h1 and h2 of the encoder section taken as the horizontal and vertical axes respectively. [Figure 12] A diagram showing an example of setting a path on the two-dimensional plane of FIG. 11. [Figure 13] A diagram showing the behavioral data obtained by reconstruction by the decoder section for the data of 10 points on the path of FIG. 12. [Figure 14] A flowchart showing the process according to the first modification example. [Figure 15] A diagram showing the path setting in the first modification example. [Figure 16] A flowchart showing the process according to the second modification example. [Figure 17] A diagram showing the path setting in the second modification example. [Figure 18] A diagram showing the path setting in the third modification example. [Figure 19] A diagram for explaining the steepest descent method. [Figure 20] A diagram showing an example of the hardware configuration of an analyzer.

Embodiments for Carrying Out the Invention

[0011] Hereinafter, an embodiment of the analytical device relating to this disclosure will be described with reference to the drawings. The analytical device relating to this disclosure performs data analysis for shifting from the first customer group to the second customer group in an environment in which there are two customer groups, the first customer group and the second customer group, each consisting of multiple customers with different characteristics. Hereinafter, as an example, the first customer group will be the customer group to which the multiple customers performing the aforementioned offline procedures belong, and will be referred to as the "offline group." The second customer group will be the customer group to which the multiple customers performing the aforementioned online procedures belong, and will be referred to as the "online group."

[0012] (Regarding the configuration and functions of the analytical instrument) As shown in Figure 1, the analysis device 10 comprises a route setting unit 11, a derivation unit 12, and a learning unit 13. The functions of each unit are outlined below.

[0013] The route setting unit 11 has the following functions in an environment where offline and online groups exist: (1) compress multidimensional vectors representing the behavioral and attribute data of each input group into two dimensions; (2) plot the two-dimensional data of each group obtained by the above compression on a two-dimensional coordinate system; and (3) set a route from feature points in the distribution of the offline group plotted on the two-dimensional coordinate system to feature points in the distribution of the online group. As functional blocks, such a route setting unit 11 includes an encoder unit 11A for realizing the function of (1) above, a plotting unit 11B for realizing the function of (2) above, and a setting unit 11C for realizing the function of (3) above.

[0014] The derivation unit 12 has the function of deriving data relating to multiple waypoints on a route set by the route setting unit 11 as data representing a gradual change from the offline group to the online group. As functional blocks, the derivation unit 12 includes a coordinate group calculation unit 12A that calculates the two-dimensional coordinates of the multiple waypoints as data relating to the multiple waypoints on the route, and a decoder unit 12B that reorganizes the data based on the calculated two-dimensional coordinates of the multiple waypoints. The decoder unit 12B, as will be described in detail later, has the function of restoring from two-dimensional variables h1 and h2, which will be described later and correspond to the calculated two-dimensional coordinates of the multiple waypoints, to a vector of the same dimension as the input multidimensional vector, and predicting whether each customer belongs to the offline group or the online group based on the above two-dimensional variables for each customer, and outputting the prediction result.

[0015] The learning unit 13, as will be described in detail later, has the function of learning the parameters of the neural networks that constitute the encoder unit 11A and the decoder unit 12B, respectively.

[0016] Figure 2 shows the processes performed in the analysis device 10. Details will be described later, but steps S1 to S5 are performed by the encoder unit 11A, plotting unit 11B, setting unit 11C, coordinate group calculation unit 12A, and decoder unit 12B, respectively. On the other hand, the learning unit 13 learns the parameters of the neural network described above, which is performed before the processes in Figure 2. Subsequently, it is executed triggered by the arrival of predetermined periodic execution timings or by a start instruction input from the operator of the analysis device 10.

[0017] The analysis device 10, configured as described above, receives various data related to customer behavior and attributes. This data will now be explained.

[0018] Figure 3 shows an example of customer behavior data. Behavioral data includes an "identifier (hereinafter referred to as "ID")" that identifies an individual customer, a "time step" in which a particular action was performed, and the specific "action." In Figure 3, the time step is represented by a discrete value (for example, a value indicating which of a predetermined time zones it was), but it may also be represented by a specific time. However, the number of time steps is a predetermined number L. In addition, "action" here can be represented by, for example, the type of web page viewed, a telephone inquiry, or a store visit. Figure 3 shows an example of data for two customers with IDs "1" and "2," but in this embodiment, it is assumed that there are N customers.

[0019] Figure 4 shows a matrix obtained by transforming the data from Figure 3 to represent the actions of each user in chronological order. In the matrix in Figure 4, "t1" represents time step 1 in Figure 3, "t2" represents time step 2, and "tL" represents time step L. The "t1" column lists the actions performed at time step 1 for each customer, associated with their ID. The same applies to the "t2"..."tL" columns.

[0020] Figure 5 shows an example of converting the elements of the matrix in Figure 4 into numerical values. In Figure 5, A=1, B=2, and C=3 are used for the numerical conversion. In Figure 5, the N×L matrix shown in the thick border, excluding the ID column, is denoted as Xt. The encoder unit 11A processes the input customer behavior data (Figure 3) through the state shown in Figure 4 into the N×L matrix Xt shown in Figure 5.

[0021] Figure 6 shows an example of data related to customer attributes. The attribute data includes an "ID" that identifies the individual customer and "variables" that represent the attributes. Here, we illustrate the case where there are M attributes from s1 to sM. For example, for attributes s1 and s2, the variables are "present (the customer has that attribute)" or "absent (the customer does not have that attribute)," and for attribute sM, the variable is a numerical value (ranging from 0 to 1) that indicates the degree to which the customer fits that attribute, such as "0.1" or "0.2."

[0022] Figure 7 shows an example of converting the non-numeric variables ("yes" and "no") in Figure 6 to numerical values, with "yes" being set to "1" and "no" to "0". In Figure 7, the N×M matrix shown in the thick border, excluding the ID column, is called Xs. The encoder unit 11A processes the input attribute data of N customers (Figure 6) into the N×M matrix Xs shown in Figure 7.

[0023] On the other hand, data that is not input to the encoder unit 11A but is input to the learning unit 13 and used for learning parameters in the neural network described later includes information, as shown in Figure 8, indicating whether each customer performed the procedure through an online or offline channel (procedure method). Figure 9 shows an example in which the data in Figure 8 has been converted to numerical values, with offline procedures being "0" and online procedures being "1". In Figure 9, the matrix shown in the thick border excluding the ID column is called Y.

[0024] (Regarding the processing performed in the analyzer 10) The following describes the processes performed in the analysis device 10, including the parameter learning process by the learning unit 13 and the series of processes shown in Figure 2.

[0025] First, we will explain the parameter learning process performed by the learning unit 13 in the multilayer neural networks that constitute the encoder unit 11A and the decoder unit 12B, respectively. Figure 10 shows the deep learning neural network used in the analysis device 10. As an example, the aforementioned matrices Xt and Xs are concatenated and input to the encoder unit 11A, which compresses them into two-dimensional latent variables h1 and h2. Then, the decoder unit 12B restores them to the same dimensions as the input and outputs the reconstructed matrix Xt* and the reconstructed matrix Xs*. Furthermore, the decoder unit 12B predicts the channel (procedure method) from the above two-dimensional latent variables and outputs the prediction result Y*. Thus, in the neural network, matrices Xt and Xs are inputs, and the reconstructed results Xt*, Xs*, and the prediction result Y* are output.

[0026] The learning unit 13 learns the parameters such that the difference between matrices Xt and Xt*, the difference between matrices Xs and Xs*, and the difference between matrices Y and Y* are minimized. That is, E=k×(|Xt*-Xt|+|Xs*-Xs|)+(1-k)×|Y*-Y| The parameters are adjusted to minimize the evaluation function E, which is represented by [formula]. For example, if Y* is expressed as Y* = w1 × h1 + w2 × h2, then w1 and w2 are parameters. k is a coefficient for adjusting the error and takes a value between 0 and 1. To improve the accuracy of reconstructing matrices Xs and Xt, increase the value of k, while to improve the prediction accuracy of matrix Y, decrease the value of k. As shown in Figure 10, by adding a network that predicts matrix Y in addition to the network that reconstructs matrices Xt and Xs, the two-dimensional latent variables h1 and h2 for various customers are more clearly separated into two groups (offline group and online group) represented by matrix Y, and as a result, the effect when a customer changes from one group to the other is expected to be more clearly shown. Furthermore, in addition to evaluating the compression of multidimensional vectors for each customer into two-dimensional variables, it becomes possible to evaluate the prediction results regarding whether each customer belongs to the offline group or the online group.

[0027] In the network shown in Figure 10, it is not necessary to input both matrices Xt and Xs to the encoder unit 11A simultaneously; either one is sufficient. Furthermore, the encoder unit 11A and decoder unit 12B do not necessarily have to be composed solely of neural networks. For example, in the encoder unit 11A, a neural network may be used to compress a multidimensional vector to an intermediate layer of three or more dimensions, and then principal component analysis may be used to compress the intermediate layer back to a two-dimensional vector. Similarly, in the decoder unit 12B, the inverse transform of principal component analysis may be used to reconstruct the two-dimensional vector to an intermediate layer of three or more dimensions, and then the neural network may be used to reconstruct the multidimensional vector from the intermediate layer. Also, when inputting matrices Xt and Xs, the matrices Xt and Xs may be pre-concatenated to form an (L+M) dimensional input, or the L-dimensional and M-dimensional inputs may be defined separately and combined within the neural network. The decoder unit 12B may also be internally divided into L-dimensional and M-dimensional intermediate layers, or it may be treated as a single (L+M) dimensional output.

[0028] Next, we will explain the series of processes performed in the analyzer 10, following the flowchart in Figure 2.

[0029] First, the encoder unit 11A converts the input data into latent variables (step S1 in Figure 2). Specifically, the encoder unit 11A processes the input behavior data of N customers (Figure 3) through the state shown in Figure 4 into the N×L matrix Xt shown in Figure 5, and processes the input attribute data of N customers (Figure 6) into the N×M matrix Xs shown in Figure 7. Then, for each individual ID, the encoder unit 11A compresses the L-dimensional vector in matrix Xt and the M-dimensional vector in matrix Xs into two dimensions, converting the multidimensional vector consisting of matrices Xt and Xs into two-dimensional latent variables h1 and h2, as shown in the upper half of the flow in Figure 10. The latent variables h1 and h2 obtained for each individual ID (each individual customer) are passed to the plotting unit 11B.

[0030] Next, the plotting unit 11B plots the 2D data (latent variables h1, h2) for each customer in a 2D coordinate system (step S2 in Figure 2). At this time, since it is possible to determine from the attribute data whether each customer belongs to the offline group or the online group, the plotting unit 11B plots the offline customers and the online customers in a way that distinguishes between them. In the example shown in Figure 11, the offline customers are represented by black circles and the online customers are represented by white circles, and the black and white circles in Figures 12, 15, 17, and 18 described later are similar. In this way, the plotting unit 11B in the route setting unit 11 displays and outputs the distribution of the offline group and the online group in the 2D coordinate system, that is, visualizes them.

[0031] Next, the setting unit 11C sets a path that starts with a feature point in the distribution of the offline group plotted in a two-dimensional coordinate system and ends with a feature point in the distribution of the online group (step S3 in Figure 2). Here, the centroid position of the distribution of each group is used as the "feature point" in the distribution of each group. This sets a path, for example, shown by arrow A in Figure 12. Although Figure 12 shows an example where the path is a straight line, the path to be set may also be a curve. In addition to using the centroid position of the distribution of each group as the "feature point" in the distribution of each group, a path may also be set that connects the two points with the longest distance between them, which is a combination of points in the offline group and points in the online group. Setting a path that connects the two points with the longest distance in this way makes it possible to increase the change in features along the path, which has the advantage of allowing a clearer difference to be obtained as a stepwise difference between the two groups.

[0032] The coordinate group calculation unit 12A then calculates the 2D coordinates of multiple (for example, 10) waypoints along the path as data (step S4 in Figure 2). The 2D coordinates of the multiple waypoints obtained through the calculation are passed to the decoder unit 12B.

[0033] Furthermore, the decoder unit 12B reorganizes the data based on the calculated 2D coordinates of the multiple passpoints (step S5 in Figure 2). In step S5, as shown in Figure 10, the decoder unit 12B restores the 2D variables h1 and h2 corresponding to the calculated 2D coordinates of each passpoint into a vector of the same dimension as the input multidimensional vector. As a result, multidimensional vector data for the number of passpoints (10 in this case) is obtained in the vertical direction of the table shown in Figure 13, that is, data on customer behavior and attributes corresponding to each of the 10 passpoints. The upward arrow at the left end of Figure 13 corresponds to arrow A on the 2D plane in Figure 12, and in Figure 13, moving from bottom to top, it shows suggestions regarding behavior and attributes when gradually changing from the offline group to the online group, that is, suggestions regarding which behavioral data and which attribute data should be changed and by how much.

[0034] Furthermore, in step S5 of Figure 2, the decoder unit 12B predicts whether each customer belongs to the offline group or the online group based on the two-dimensional variables for each customer, and outputs the prediction result Y* shown in Figure 10. This prediction result Y*, along with the reconstruction results Xt* and Xs* mentioned above, is used by the learning unit 13 to learn the network parameters.

[0035] Through the embodiments described above, it becomes possible to obtain insights into which behavioral data and attribute data should be changed and to what extent as customers gradually transition from the offline group to the online group, thereby deriving highly useful data for encouraging customers to change their behavior from the offline group to the online group.

[0036] Furthermore, as shown in Figure 10, by adding a network that predicts matrix Y and outputs the prediction result Y*, in addition to the network that reconstructs matrices Xt and Xs, the two-dimensional latent variables h1 and h2 for various customers can be more clearly separated into two groups (offline group and online group) represented by matrix Y. As a result, the effect of customers changing from the offline group to the online group is expected to be more clearly expressed. In addition to evaluating the compression of multidimensional vectors for each customer into two-dimensional variables, it becomes possible to evaluate the prediction results regarding whether each customer belongs to the offline group or the online group.

[0037] Furthermore, the plotting unit 11B in the route setting unit 11 displays and outputs the distribution of offline groups and online groups in a two-dimensional coordinate system, i.e., visualizes them. As a result, operators of the analysis device 10 can easily visually grasp the distribution and routes of each group, and can smoothly perform fine adjustments to the routes as needed.

[0038] (Regarding various variations) In the following, we will sequentially describe two modifications of the above embodiment: Modification 1 and 2, which involve reconstructing the system with certain constraints, and Modification 3, which involves determining multiple contour lines in a two-dimensional coordinate system by connecting points where the ratio of the first distribution to the second distribution is equal, and then setting a path based on these contour lines.

[0039] (Variation 1) Even when trying to change customers' behavior or attributes, there are some things that are not practically easy to change, such as their address (current place of residence). Taking these circumstances into consideration, Modification 1 shows an example in which the route setting unit 11 in the analysis device 10 obtains the distribution of the offline group and the distribution of the online group under the constraint condition that a certain element (such as an address) is fixed, and sets a route within the range of both obtained distributions, that is, an example of route setting that satisfies the constraint condition of fixing a certain element.

[0040] In the first modified example, the process shown in Figure 14 is performed in the analyzer 10. Since steps S3A and S3B differ from the process shown in Figure 2 described above, these steps will be explained. In step S3A of Figure 14, the plotting unit 11B plots data that satisfies the constraint condition of fixing a certain element (for example, an address), in addition to the plotting in step S2. Then, in step S3B, the setting unit 11C sets a path within the range of the area plotted in step S3A. For example, in the two-dimensional plane shown in Figure 15, if the area plotted in step S3A (the range of both distributions) is the elliptical area R1, the setting unit 11C sets a path within the range of area R1 as shown by arrow A1.

[0041] According to this modified example 1, it is possible to appropriately set a path under the constraint of fixing a certain element.

[0042] (Modification 2) On the other hand, when changing customer behavior or attributes, there are certain elements that must be avoided. Taking these circumstances into consideration, Modification 2 shows an example in which the route setting unit 11 in the analysis device 10 obtains the distribution of offline groups and the distribution of online groups from data consisting of elements to be avoided, and sets a route that avoids the range of both obtained distributions, that is, an example of setting a route that satisfies the constraint of avoiding a certain element.

[0043] In the modified example 2, the process shown in Figure 16 is performed in the analyzer 10. Since steps S3A and S3C differ from the process shown in Figure 2 described above, these steps will be explained. In step S3A of Figure 16, the plotting unit 11B plots data that corresponds to the element to be avoided, in addition to the plotting in step S2. Then, in step S3C, the setting unit 11C sets a path to avoid the area plotted in step S3A. For example, in the two-dimensional plane shown in Figure 17, if the area plotted in step S3A (the range of both distributions) is the elliptical area R2, the setting unit 11C sets a path to avoid area R2, for example, as shown by arrow A2.

[0044] According to this modified example 2, a suitable path can be set under the constraint of avoiding a certain element.

[0045] (Variation 3) In Modification 3, the path setting unit 11 of the analyzer 10 calculates multiple lines (referred to as "contour lines" in this case) in a two-dimensional coordinate system, connecting points where the ratio of the distribution of the offline group to the distribution of the online group is equal, according to the ratio, and sets a path based on the multiple contour lines obtained.

[0046] Specifically, as shown in Figure 18, the route setting unit 11 uses a certain point as a reference in a two-dimensional coordinate system and connects points in the surrounding point cloud where the ratio of "points belonging to the offline group" to "points belonging to the online group" is equal. This unit then obtains multiple contour lines shown as dashed lines in Figure 18, and based on the obtained contour lines, sets route A3 using one of the following setting methods 1 to 3.

[0047] As setting method 1, the route setting unit 11 may adopt a method of setting a rapidly changing route by taking a route with narrowly spaced contour lines. In this case, the advantage is that the difference to be changed becomes larger, making it possible to clearly define the difference to be changed.

[0048] As a second setting method, the route setting unit 11 may adopt a method of setting a gradually changing route by choosing a route with wide spacing between contour lines. In this case, since the difference that needs to be changed becomes smaller, there is an advantage in that it is possible to encourage changes that are less burdensome for the customer group (in this case, the offline customer group).

[0049] As a third setting method, the route setting unit 11 may adopt a method of exploratory route setting based on the gradient at a certain point, similar to the steepest descent method. In this case, since the route is set automatically, there is an advantage in that a human (for example, the operator of the analysis device 10) can clearly identify the differences that should be changed compared to the route set by looking at the contour lines visualized in Figure 18.

[0050] Here, we will explain an example of exploratory path setting based on the gradient at a certain point using the steepest descent method. From the coordinates (h1(t), h2(t)) of a certain point at a certain time step t, the coordinates (h1(t+1), h2(t+1)) of the point to be reached at the next time (t+1) are expressed by the following equation (1).

number

[0051] In the upper right of Figure 19, an example of the density distribution around a certain point (h1(t), h2(t)) (in this case, a 3x3 surrounding area) is shown, with the aforementioned point (h1(t), h2(t)) located in the central area. In this density distribution example, both the h1 direction and the h2 direction are Δh i =2, and ΔD1=4-2, ΔD2=5-1. Here, if we set the parameter d=2, then by substituting this into equation (1) above, h1(t+1)=h1(t)-2×(4-2) / 2=h1(t)-2 h2(t+1)=h2(t)-2×(5-1) / 2=h2(t)-4 This results in a path that moves "-2" in the h1 direction (left-right) and "-4" in the h2 direction (up-down). This method using the steepest descent method can be adopted as one variation of the path setting method (here, setting method 3).

[0052] The gist of this disclosure is found in the following [1] to

[10] . [1] In an environment in which a first customer group includes multiple customers having a first characteristic and a second customer group includes multiple customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the same two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit that derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, An analytical device equipped with the following features. [2] The route setting unit is: The analytical apparatus according to [1], which sets a path starting from the centroid position of the first distribution and ending from the centroid position of the second distribution. [3] The route setting unit, The analytical apparatus according to [1], which sets a path connecting the two points that have the longest distance between them, using a combination of points from the first distribution and points from the second distribution. [4] The route setting unit, The analytical apparatus according to [1], which obtains multiple contour lines corresponding to the ratio obtained by connecting points in the two-dimensional coordinate system where the ratio of the first distribution and the second distribution are equal, and sets a path that minimizes the distance between contour lines. [5] The route setting unit, The analytical apparatus according to [1], which obtains multiple contour lines corresponding to the ratio obtained by connecting points in the two-dimensional coordinate system where the ratio of the first distribution and the second distribution are equal, and sets a path that maximizes the distance between contour lines. [6] The route setting unit, The analytical apparatus according to [1], which obtains multiple contour lines corresponding to the ratio by connecting points in the two-dimensional coordinate system where the ratio of the first distribution and the second distribution are equal, and sets a path based on the gradient of the ratio at a certain point. [7] The route setting unit, An analytical apparatus according to any one of [1] to [6], which obtains the first distribution and the second distribution under the constraint condition of fixing a certain element, and sets the path within the range of the obtained first distribution and second distribution. [8] The route setting unit, An analytical apparatus according to any one of [1] to [6], which obtains the first distribution and the second distribution from data consisting of elements to be avoided, and sets the path so as to avoid the ranges of the obtained first distribution and the second distribution. [9] The analysis apparatus according to any one of [1] to [8], wherein the route setting unit displays and outputs the first distribution and the second distribution in the two-dimensional coordinate system.

[10] The route setting unit, It includes an encoder unit that compresses the multidimensional vectors for each customer in the first customer group and the second customer group into two-dimensional variables, The aforementioned derivation section is, It includes a decoder unit that reconstructs the two-dimensional variables for each customer, corresponding to data about multiple waypoints along the aforementioned path, into a vector of the same dimension as the multi-dimensional vector, The decoder unit further predicts, based on the two-dimensional variables for each customer, whether the customer belongs to the first customer group or the second customer group, and outputs the prediction result, as described in any one of [1] to [9].

[0053] (Explanation of terms, explanation of hardware configuration (Figure 20), etc.) The block diagrams used in the description of the above embodiments show functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method of realizing each functional block is not particularly limited. That is, each functional block may be realized using one device that is physically or logically coupled, or it may be realized using two or more physically or logically separated devices that are directly or indirectly connected (for example, using wired or wireless connections). A functional block may also be realized by combining the above one device or the above multiple devices with software.

[0054] Functions include, but are not limited to, judgment, decision, judgment, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, resolution, selection, selection, establishment, comparison, assumption, expectation, assumption, broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating (mapping), and assigning. For example, a functional block (configuration part) that enables transmission is called a transmitting unit or transmitter. As mentioned above, the method of implementation is not particularly limited.

[0055] For example, an analytical apparatus in one embodiment of the present disclosure may function as a computer that performs the processing of the present disclosure. Figure 20 is a diagram showing an example of the hardware configuration of an analytical apparatus 10 according to one embodiment of the present disclosure. The analytical apparatus 10 described above may be physically configured as a computer device including a processor 1001, memory 1002, storage 1003, communication device 1004, input device 1005, output device 1006, bus 1007, etc.

[0056] In the following explanation, the term "device" can be replaced with "circuit," "device," "unit," etc. The hardware configuration of the analytical apparatus 10 may include one or more of the devices shown in the figure, or it may be configured to omit some of the devices.

[0057] Each function in the analysis device 10 is realized by loading predetermined software (programs) onto hardware such as the processor 1001 and memory 1002, which allows the processor 1001 to perform calculations, control communication by the communication device 1004, and control at least one of data reading and writing in the memory 1002 and storage 1003.

[0058] The processor 1001 controls the entire computer, for example, by running an operating system. The processor 1001 may consist of a central processing unit (CPU) that includes interfaces with peripheral devices, control units, arithmetic units, registers, and so on.

[0059] Furthermore, the processor 1001 reads programs (program code), software modules, data, etc., from at least one of the storage 1003 and the communication device 1004 into the memory 1002 and executes various processes accordingly. The program used is one that causes the computer to execute at least a part of the operations described in the above embodiment. Although the above processes have been described as being executed by one processor 1001, they may be executed simultaneously or sequentially by two or more processors 1001. The processor 1001 may be implemented by one or more chips. The program may also be transmitted from a network via a telecommunications line.

[0060] Memory 1002 is a computer-readable recording medium and may consist of at least one of the following: ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), RAM (Random Access Memory), etc. Memory 1002 may also be called a register, cache, main memory, etc. Memory 1002 can store executable programs (program code), software modules, etc., for carrying out a wireless communication method according to one embodiment of the present disclosure.

[0061] Storage 1003 is a computer-readable recording medium and may consist of at least one of the following: 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 multipurpose disc, a Blu-ray® disc), a smart card, flash memory (e.g., a card, a stick, a key drive), a floppy® disk, a magnetic strip, etc. Storage 1003 may also be called an auxiliary storage device. The above-mentioned storage medium may be, for example, a database, server, or other suitable medium including at least one of memory 1002 and storage 1003.

[0062] The communication device 1004 is hardware (transceiver / receiver device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as a network device, network controller, network card, communication module, etc. The communication device 1004 may be configured to include, for example, a high-frequency switch, duplexer, filter, frequency synthesizer, etc., in order to implement at least one of frequency division duplex (FDD) and time division duplex (TDD).

[0063] The input device 1005 is an input device that accepts input from an external source (e.g., a keyboard, mouse, microphone, switch, button, sensor, etc.). The output device 1006 is an output device that outputs to an external source (e.g., a display, speaker, LED lamp, etc.). The input device 1005 and the output device 1006 may be configured as an integrated unit (e.g., a touch panel).

[0064] Furthermore, each device, such as the processor 1001 and memory 1002, is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or different buses may be configured for each device.

[0065] Furthermore, the analysis device 10 may be configured to include hardware such as a microprocessor, a digital signal processor (DSP), an ASIC (Application Specific Integrated Circuit), a PLD (Programmable Logic Device), and an FPGA (Field Programmable Gate Array), and some or all of each functional block may be realized by such hardware. For example, the processor 1001 may be implemented using at least one of these hardware components.

[0066] Information notification is not limited to the embodiments described herein and may be carried out by other means. For example, information notification may be carried out by physical layer signaling (e.g., DCI (Downlink Control Information), UCI (Uplink Control Information)), upper layer signaling (e.g., RRC (Radio Resource Control) signaling, MAC (Medium Access Control) signaling, broadcast information (MIB (Master Information Block), SIB (System Information Block))), other signals, or combinations thereof. RRC signaling may also be called RRC messages, and may be, for example, RRC Connection Setup messages, RRC Connection Reconfiguration messages, etc.

[0067] Each aspect / embodiment described in this disclosure includes LTE (Long Term Evolution), LTE-A (LTE-Advanced), SUPER 3G, IMT-Advanced, 4G (4th generation mobile communication system), 5G (5th generation mobile communication system), 6th generation mobile communication system (6G), xth generation mobile communication system (xG) (xG (where x is, for example, an integer or decimal)), FRA (Future Radio Access), NR (new Radio), New radio access (NX), Future generation radio access (FX), W-CDMA (registered trademark), GSM (registered trademark), CDMA2000, UMB (Ultra Mobile Broadband), IEEE 802.11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), and IEEE This may apply to at least one system utilizing 802.20, UWB (Ultra-WideBand), Bluetooth®, or other appropriate systems, and to next-generation systems extended, modified, created, or defined based thereon. It may also apply to a combination of multiple systems (for example, a combination of at least one of LTE and LTE-A with 5G).

[0068] The processing procedures, sequences, flowcharts, etc., of each aspect / embodiment described herein may be reordered, provided they are consistent with each other. For example, the methods described herein present various step elements in an exemplary order and are not limited to that specific order.

[0069] Input and output information may be stored in a specific location (e.g., memory) or managed using a management table. Input and output information may be overwritten, updated, or appended to. Output information may be deleted. Input information may be transmitted to other devices.

[0070] The determination may be made by a value represented by 1 bit (0 or 1), by a boolean value (true or false), or by a numerical comparison (for example, a comparison with a predetermined value).

[0071] Each aspect / embodiment described herein may be used individually, in combination, or switched between as needed during implementation. Furthermore, notification of specific information (e.g., notification that "X is") is not limited to explicit notification, but may also be implicit (e.g., by not providing such notification).

[0072] Although the present disclosure has been described in detail above, it will be clear to those skilled in the art that the present disclosure is not limited to the embodiments described herein. The present disclosure can be implemented in modified and altered forms without departing from the intent and scope of the present disclosure as defined by the claims. Therefore, the descriptions in the present disclosure are illustrative and not intended to be restrictive in any way.

[0073] Software should be broadly interpreted to mean instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, applications, software applications, software packages, routines, subroutines, objects, executable files, execution threads, procedures, functions, and so on, whether they are called software, firmware, middleware, microcode, hardware description languages, or by any other name.

[0074] Furthermore, software, instructions, information, etc., may be transmitted and received via a transmission medium. For example, if software is transmitted from a website, server, or other remote source using at least one of wired technology (such as coaxial cable, fiber optic cable, twisted pair, or digital subscriber line (DSL)) and wireless technology (such as infrared or microwave), then at least one of these wired and wireless technologies is included in the definition of a transmission medium.

[0075] The information, signals, etc. described in this disclosure may be represented using any of the various different techniques. For example, the data, instructions, commands, information, signals, bits, symbols, chips, etc. that may be referred to throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or magnetic particles, optical fields or photons, or any combination thereof.

[0076] In addition, terms used in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, at least one of the communication channel and the symbol may be a signal (signaling). Also, the signal may be a message. Furthermore, the component carrier (CC) may be called a carrier frequency, cell, frequency carrier, etc.

[0077] The terms “system” and “network” as used in this disclosure are interchangeable.

[0078] Furthermore, the information, parameters, etc., described in this disclosure may be expressed using absolute values, relative values ​​from a given value, or corresponding other information. For example, wireless resources may be indicated by an index.

[0079] The names used for the parameters described above are not restrictive in any way. Furthermore, the formulas and other expressions using these parameters may differ from those expressly disclosed in this disclosure. Various communication channels (e.g., PUCCH, PDCCH, etc.) and information elements can be identified by any suitable name, and therefore, the various names assigned to these various communication channels and information elements are not restrictive in any way.

[0080] As used in this disclosure, the terms “determining” and “determining” may encompass a wide variety of actions. “Determining” may include, for example, judging, calculating, computing, processing, deriving, investigating, looking up, searching, inquiry (e.g., searching in a table, database, or other data structure), and ascertaining. “Determining” may also include, for example, receiving (e.g., receiving information), transmitting (e.g., sending information), input, output, and accessing (e.g., accessing data in memory). Furthermore, "judgment" and "decision" can include considering something as having been "judged" or "decided" after resolving, selecting, choosing, establishing, comparing, etc. In other words, "judgment" and "decision" can include considering something as having been "judged" or "decided" after some action. Also, "judgment (decision)" can be reinterpreted as "assuming," "expecting," or "considering."

[0081] In this disclosure, the phrase "based on" does not mean "based solely on" unless otherwise specified. In other words, the phrase "based on" means both "based solely on" and "based at least on."

[0082] Any reference to elements using the designations “first,” “second,” etc., as used in this disclosure does not generally limit the quantity or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Accordingly, references to the first and second elements do not imply that only two elements may be employed, or that the first element must precede the second element in any way.

[0083] Where the terms “include,” “including,” and variations thereof are used in this disclosure, these terms are intended to be inclusive, as is the term “comprising.” Furthermore, the term “or” as used in this disclosure is not intended to mean exclusive OR.

[0084] In this disclosure, if articles are added through translation, such as a, an, and the in English, this disclosure may include the fact that the noun following these articles is plural.

[0085] In this disclosure, the term "A and B are different" may mean "A and B are different from each other." The term may also mean "A and B are each different from C." Terms such as "separate" and "combine" may be interpreted similarly to "different." [Explanation of symbols]

[0086] 10...Analysis device, 11...Route setting unit, 11A...Encoder unit, 11B...Plotting unit, 11C...Setting unit, 12...Derivation unit, 12A...Coordinate group calculation unit, 12B...Decoder unit, 13...Learning unit, 1001...Processor, 1002...Memory, 1003...Storage, 1004...Communication device, 1005...Input device, 1006...Output device, 1007...Bus.

Claims

1. In an environment where there exists a first customer group consisting of multiple customers having a first characteristic and a second customer group consisting of multiple customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the same two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, A path is set with the centroid of the first distribution as the starting point and the centroid of the second distribution as the ending point. Analyzer.

2. In an environment in which a first customer group comprising a plurality of customers having a first characteristic and a second customer group comprising a plurality of customers having a second characteristic different from the first characteristic exist, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, A path is set that connects the two points with the longest distance between them, using the combination of points from the first distribution and points from the second distribution. Analyzer.

3. In an environment in which a first customer group includes a plurality of customers having a first characteristic and a second customer group includes a plurality of customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, In the aforementioned two-dimensional coordinate system, multiple contour lines corresponding to the ratio are obtained by connecting points where the ratio of the first distribution and the second distribution are equal, and a path is set that minimizes the distance between the contour lines. Analyzer.

4. In an environment in which a first customer group includes a plurality of customers having a first characteristic and a second customer group includes a plurality of customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, using a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, In the aforementioned two-dimensional coordinate system, multiple contour lines corresponding to the ratio are obtained by connecting points where the ratio of the first distribution and the second distribution are equal, and a path is set that maximizes the distance between the contour lines. Analyzer.

5. In an environment in which a first customer group comprising a plurality of customers having a first characteristic and a second customer group comprising a plurality of customers having a second characteristic different from the first characteristic exist, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, In the aforementioned two-dimensional coordinate system, multiple contour lines corresponding to the ratio are obtained by connecting points where the ratio of the first distribution and the second distribution are equal, and a path is set based on the gradient of the ratio at a certain point. Analyzer.

6. In an environment in which a first customer group includes a plurality of customers having a first characteristic and a second customer group includes a plurality of customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, based on a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, Under the constraint of fixing a certain element, the first distribution and the second distribution are obtained, and the path is set within the range of the obtained first distribution and second distribution. Analyzer.

7. In an environment in which a first customer group includes a plurality of customers having a first characteristic and a second customer group includes a plurality of customers having a second characteristic different from the first characteristic, a path setting unit sets a path from a feature point in the first distribution to a feature point in the second distribution, using a first distribution in a two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the first customer group into two dimensions, and a second distribution in the two-dimensional coordinate system obtained by compressing a multidimensional vector representing the behavioral data and attribute data of the second customer group into two dimensions. A derivation unit derives data relating to multiple waypoints on the route set by the route setting unit as data representing a gradual change from the first customer group to the second customer group, Equipped with, The aforementioned route setting unit, Obtain the first and second distributions from data consisting of elements to be avoided, and set the path so as to avoid the ranges of the obtained first and second distributions. Analyzer.

8. The route setting unit displays and outputs the first distribution and the second distribution in the two-dimensional coordinate system. The analytical apparatus according to any one of claims 1 to 7.

9. The aforementioned route setting unit, It includes an encoder unit that compresses the multidimensional vectors for each customer in the first customer group and the second customer group into two-dimensional variables, The aforementioned derivation section is, It includes a decoder unit that reconstructs the two-dimensional variables for each customer, corresponding to data about multiple waypoints along the aforementioned route, into a vector of the same dimension as the multi-dimensional vector, The decoder unit further predicts, based on the two-dimensional variables for each customer, whether the customer belongs to the first customer group or the second customer group, and outputs the prediction result. The analytical apparatus according to any one of claims 1 to 7.

Images