JP2016095626A - Estimation device and estimation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable the trend of a customer in a market to be grasped.SOLUTION: An estimation device comprises: a conversion section which converts time series data on each of the number of inflow persons, the number of outflow persons, and a net increase number of persons to be employed about the first competitive element into time series data with a time axis as log; a generation section which generates a broken-line model approximate to each converted time series data; and a first estimation section which estimates a change amount of inclination of the broken-line model related to the number of inflow persons or the broken-line model related to the number of outflow persons in a period of time not corresponding to a change point of the broken-line model related to the net increase number, as a reflow amount which does not affect the net increase number but is a component changing the number of inflow persons or the number of outflow persons.SELECTED DRAWING: Figure 2

Description

  The present invention relates to an estimation apparatus and an estimation method.

  In the product and service market, it is common for products or services of a plurality of operators to compete and to compete for market share. In such a market, when changes occur in the market share of the operator, this is detected, and the change in the situation changes such as what kind of interaction or influence relationship has occurred between the market share of the business operator. If we can understand the structure, and if we can analyze and evaluate the impact of changes in the situation, such as how the resulting impact relationship can bring about future market occupancy, we will be able to implement market measures and Judgment of change can be made more accurately and quickly. Thus, if there is a way to quantitatively estimate the structure of market changes and past and future impacts, chances of lost opportunities and investment failures will increase, and economic Therefore, it is considered useful.

  In recent years, prediction of market growth has been started in order to achieve the above object (see, for example, Non-Patent Document 1 and Patent Document 1).

Japanese Patent No. 47777941

S. Shimogawa, M. Shinno, and H. Saito, Structure of S-shaped growth in innovation diffusion, Physical Review E 85, 056121 (2012).

  Until now, the technology has been able to analyze the structure of the change in the situation, but when the situation changed, whether the customer who currently contracted their services canceled and moved to another company (impact on the outflow), Established a method to determine the movement of customers between services including inflow and outflow, whether the customer who planned to join the company contracted with another company (impact on inflow) due to a certain opportunity (for example, cost merit by campaign) It has not been. In addition, from the perspective of the business owner, there is no impact on the net increase in the number of customers, but no method has been established for obtaining a component that changes the number of inflows and outflows (hereinafter referred to as “reflux”).

  Reflux is a situation that occurs in highly competitive services. Specifically, when customers benefit from changing their services regularly (for example, mobile phones) In the market, there is a situation where customers who regularly inflow or outflow exist in a market, for example, when the provider is changed continuously for the purpose of cashback by changing the provider.

  The present invention has been made in view of the above points, and an object thereof is to make it possible to grasp the movement of customers in the market.

  Therefore, in order to solve the above-mentioned problem, the estimation device converts the time series data regarding each of the number of inflows, outflows, and net increase of the employer regarding the first competitive element into time series data with a logarithmic time axis. , A generation unit that generates a line model approximating each time series data after conversion, and a line model or the number of outflows related to the number of inflows at a time not corresponding to a change point of the line model related to the net increase And a first estimation unit that estimates the amount of change in the slope of the line model as a recirculation amount that does not affect the net increase, but is a component that changes the number of inflows or outflows.

  It is possible to grasp the movement of customers in the market.

It is a figure which shows the hardware structural example of the estimation apparatus in embodiment of this invention. It is a figure which shows the function structural example of the estimation apparatus in embodiment of this invention. It is a figure which shows the structural example of a contractor number memory | storage part. It is a figure which shows the structural example of an inflow / outflow number memory | storage part. It is a flowchart for demonstrating an example of the process sequence which a broken line model production | generation part performs. It is a flowchart for demonstrating an example of the process sequence which a recirculation | reflux amount estimation part performs. It is a flowchart for demonstrating an example of the process sequence which an influence amount estimation part performs.

  In the present embodiment, in the market for services or products for consumers (hereinafter referred to as “service”), competing factors such as service providers compete for customer acquisition. The number of customer fluctuations between competing services and the component that changes the number of inflows and outflows (hereinafter referred to as “reflux”), although not affecting the net increase in the number of customers from the perspective of the business owner, are estimated. The estimation device 10 that performs the above will be described.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a hardware configuration example of an estimation apparatus according to an embodiment of the present invention. The estimation device 10 in FIG. 1 includes a drive device 100, an auxiliary storage device 102, a memory device 103, a CPU 104, an interface device 105, and the like that are mutually connected by a bus B.

  A program that realizes processing in the estimation apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 storing the program is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. However, the program need not be installed from the recording medium 101 and may be downloaded from another computer via a network. The auxiliary storage device 102 stores the installed program and also stores necessary files and data.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The CPU 104 executes a function related to the estimation device 10 according to a program stored in the memory device 103. The interface device 105 is used as an interface for connecting to a network.

  FIG. 2 is a diagram illustrating a functional configuration example of the estimation device according to the embodiment of the present invention. 2, the estimation device 10 includes a log origin setting unit 11, a log time conversion unit 12, a polygonal line model generation unit 13, a reflux amount estimation unit 14, an influence amount estimation unit 15, and the like. Each of these units is realized by processing that one or more programs installed in the estimation apparatus 10 cause the CPU 104 to execute. The estimation apparatus 10 also uses an inflow / outflow number storage unit 121, a contractor number storage unit 122, and the like. The inflow / outflow number storage unit 121 and the contractor number storage unit 122 can be realized using a storage device or the like that can be connected to the auxiliary storage device 102 or the estimation device 10 via a network.

The contractor number storage unit 122 stores time-series data of the number of contractors (the number of employers) of a certain service. For example, the number of contractors storage unit 122 stores the number of contractors N _j , j = 1, 2,... For each service (for example, FLET'S Hikari (registered trademark), au Hikari (registered trademark), etc.) competing with each other. Time-series data D _j = {(t _k , N _j (t _k ), k = 1, 2,..., A for a time interval [T ₀ , T ₁ ] up to a certain time in the past or in the past ., A)} is stored. Here, j is a subscript for distinguishing the service, and k is a subscript for distinguishing the time (for example, month if monthly).

  FIG. 3 is a diagram illustrating a configuration example of the contractor number storage unit. In FIG. 3, the contractor number storage unit 122 stores the number of contractors for each service for each time period.

  As for the time series data of the number of contractors for each competing service (including in-house and all other companies), the fixed broadband service is taken as an example of the “Number of telecommunications service contracts provided by the Ministry of Internal Affairs and Communications. "Share data" may be used.

The inflow / outflow number storage unit 121 stores time-series data of the number of new contractors (inflows), the number of new churners (outflows), and the number of net increasers regarding their services. The net increase in number can be calculated from the difference for each fixed period of the time-series data of the number of contractors. The net increase number I _{[t0, t1]} in a certain time interval [t ₀ , t ₁ ] is calculated by I _{[t0, t1]} = D _j (t ₁ ) −D _j (t ₀ ).

  FIG. 4 is a diagram illustrating a configuration example of the inflow / outflow number storage unit. In FIG. 4, the inflow / outflow number storage unit 121 stores the inflow number, outflow number, and net increase number for each service for each company.

  Note that values managed by the company system may be used for the time series data of the number of inflows of the company service and the time series data of the number of outflows. Further, in the present embodiment, “in-house” refers to a business entity that uses the estimation device 10. “Other companies” refers to businesses that provide services that compete with services provided by the company.

  The logarithmic origin setting unit 11 sets the logarithmic time origin from each time series data stored in the contractor number storage unit 122 or the inflow / outflow number storage unit 121. For example, paragraph 0035 of Patent Document 1 may be referred to for setting the logarithmic time origin.

The logarithmic time conversion unit 12 executes a process of converting the time axis of each time series data stored in the contractor number storage unit 122 or the inflow / outflow number storage unit 121 into a logarithm. For example, the time-series data of the number of contractors in the contractor number storage unit 122 is
D = {N (t _i ), t _i , i = 1, 2,..., N}
In this case, the logarithmic time conversion unit 12 responds to the data DL = {N (t _i ), log (t _i ), i = 1, 2,. }
Is calculated.

  For example, Patent Document 1 may be referred to for the process of converting the time axis of the time series data into a logarithm.

  The polygonal line model generation unit 13 approximates the time series data of the cumulative number of contractors, the cumulative inflow number, the cumulative outflow number, and the cumulative net increase after the time axis conversion by the logarithmic time conversion unit 12. Generate a line model.

  The reflux amount estimation unit 14 estimates an increase / decrease amount of the reflux amount in each year based on the change amount of the slope at the change point (break point) of the inflow number line model or the outflow number line model. Based on the amount of increase / decrease in the amount of reflux, the amount of reflux at a certain time of each year is estimated.

  The influence amount estimation unit 15 is based on a line model of the time series data of the net increase related to the company's service, a line model of the number of inflows or a line model of the number of outflows, and a line model of the time series data of the number of contractors related to services of other companies. Thus, we estimate the services of other companies that have affected the number of inflows or outflows related to in-house services, and the amount of the impact.

  Hereinafter, the process procedure which the estimation apparatus 10 performs is demonstrated. FIG. 5 is a flowchart for explaining an example of a processing procedure executed by the polygonal line model generation unit. The processing procedure of FIG. 5 includes the time series data of the number of contractors of each company and other companies, the time series data of the number of inflows related to the company, and the number of outflows related to the company after the time is logarithmized by the log time conversion unit 12. It is executed for each of the time series data and the time series data of the net increase related to the company.

In step S <b> 101, the polygonal line model generation unit 13 executes an approximation procedure using straight lines for each section obtained by dividing the designated section by the folding position candidates designated in advance by the user for the time series data to be processed. The designated section is any section in the time section [T ₀ , T ₁ ] and is a section designated in advance by the user.

  Subsequently, the broken line model generation unit 13 does not necessarily fold at the fold position candidates, and therefore, for each fold position candidate, evaluates whether or not the difference in inclination has changed beyond a predetermined inclination threshold at several points before and after. If it is equal to or greater than the threshold, the fold position candidate is set as a break point (S102). Therefore, a break position candidate that is less than the threshold value is not a break point.

  Subsequently, the broken line model generation unit 13 sets each broken point obtained in step S102 as an initial broken point, and generates a regression line using a known linear regression method for each broken point section (S103). . As a result, a line model is obtained for the time-series data to be processed.

  Subsequently, the broken line model generation unit 13 calculates a mean square error between the obtained broken line model and the time-series data to be processed, and sets the sum of the mean square differences as an error (S104). When the error exceeds a preset threshold value 1 (Th1) (YES in S105), the polygonal line model generation unit 13 determines a section having the maximum mean square error for each section (hereinafter referred to as “maximum section”). Introduce new division points. Specifically, the polygonal line model generation unit 13 moves the candidate for the new segment point from the start point of the maximum section to the end point of the maximum section with the width of the step value δ designated in advance by the user (S106). ), A regression line is generated for each of the two sections obtained by dividing the maximum section by the new section points (S107), and a new section point having the smallest sum of mean square errors is added to the breakpoint (S108). That is, steps S106 and S107 are repeatedly executed while the new division point is moved.

  If the width of the minimum segment among the segment segments of the polygonal line model is less than the predetermined threshold value 2 (Th2) as a result of the addition of the polygonal points (YES in S109), the polygonal line model generation unit 13 generates the polygonal line model. (S110), the process of FIG. 5 is terminated. If the width of the minimum section is equal to or greater than the predetermined threshold 2 (Th2) (NO in S109), the polygonal line model generation unit 13 repeats step S103 and subsequent steps.

  When the error calculated in S104 is equal to or less than the threshold value 1 (Th1) (NO in S105), the broken line model generation unit 13 uses, for example, the coordinate value of each broken point as information for specifying the generated broken line model. And the slope of each section are output (S111). The output value is stored in the memory device 103 or the auxiliary storage device 102, for example.

  In addition, changes in the slope (change in speed) of the time series data line model of the number of in-house services inflow and the time series data of the number of out-flows of in-house services increase or decrease the market share of the in-house service market. It is a numerical value that characterizes the speed.

  Subsequently, a processing procedure executed by the reflux amount estimation unit 14 will be described. FIG. 6 is a flowchart for explaining an example of a processing procedure executed by the reflux amount estimation unit. The process of FIG. 6 may be started in response to an instruction from the user, for example, or may be executed following the process of FIG.

  In step S201, the recirculation amount estimation unit 14 receives an input of a recirculation time and a recirculation initial year from the user. The reflux time is a time when the user determines that the influence of reflux can be observed remarkably. In this embodiment, since each time series data is monthly, a specific month (for example, April at the beginning of the year) is input as the return time. The initial year of reflux is the year before the year in which the calculation of the amount of reflux is started. For example, the year when the user determines that competition has not intensified (the year in which the amount of reflux can be regarded as 0) is input. The

  Subsequently, the recirculation amount estimation unit 14 selects data to be analyzed in calculating the recirculation amount from the time series data of the inflow number and the time series data of the outflow number (S202). Specifically, the time series data generated by the line model generation unit 13 is selected from the time series data of the number of inflows and the time series data of the number of outflows. A broken line model that is stable in time series is a broken line model that has a relatively small error (for example, the sum of mean square errors) with respect to time series data. The reason why the one that is stable in time series is selected is to select a polygonal line model indicating a state close to reality and obtain an analysis result having a relatively small error relative to the actual situation.

For example, the data with the fitting error of the polygonal line model generated by the polygonal line model generation unit 13 being within the threshold value and having a relatively small fitting error may be selected. Specifically, the determination coefficient (R ² ) generated in the regression analysis executed when the line model generation unit 13 generates the line model of the time series data of the inflow number and the line model of the time series data of the outflow number. May be selected such that the determination coefficient is equal to or greater than the threshold ε and the determination coefficient is relatively large, as in the following equation (1).

但し、他の方法によって、分析対象とするデータが選択されてもよい。以下、選択された方のデータを、「選択データ」という。

However, data to be analyzed may be selected by other methods. Hereinafter, the selected data is referred to as “selected data”.

  When there is no data corresponding to the selection condition, for example, when there is no data whose determination coefficient is equal to or greater than the threshold ε (NO in S203), the reflux amount estimation unit 14 determines that the influence amount cannot be calculated, and FIG. The process ends.

  When any one of the time series data of the number of inflows and the time series data of the number of outflows is selected as the selection data (YES in S203), the reflux amount estimation unit 14 selects the year of analysis (hereinafter, “analysis target”). Year ") is set to the year following the initial reflux year (S204).

  Subsequently, the reflux amount estimation unit 14 determines whether there is a change point (break point) in the line model of the time series data of the net increase (whether there is a change in slope) in the month corresponding to the return time of the year to be analyzed. That is, it is determined whether or not the rate of net increase has changed (S205). If there is a change point in the line model of the time series data of the net increase at the return time of the analysis target year (YES in S205), the return amount estimation unit 14 sets the increase / decrease amount of the return amount of the analysis year as a default value. Then, the increase / decrease amount (number of persons / year) of the reflux amount set for the previous year of the analysis target year is set (S206). Note that the change in the rate of the net increase can occur due to, for example, a competitive service offensive (for example, a discount measure). The increase / decrease amount of the reflux amount in the initial reflux year is zero.

  On the other hand, when there is no change point in the line model of the time series data of the net increase at the reflux time of the year to be analyzed (NO in S205), the reflux amount estimation unit 14 relates to the speed at the reflux time with respect to the line model of the selected data ( (Slope) change amount is calculated, and the change amount is set as an increase / decrease amount of the reflux amount with respect to the analysis target year (S207). The amount of change is calculated by the speed (slope) immediately after the change point−the speed (slope) immediately before the change point.

  That is, when the number of inflows or outflows changes at a time when the net increase has not changed, the factor of the change does not affect the net increase, but is a component that changes the inflow or outflow. It can be estimated that. Therefore, it can be estimated that the amount of change in the speed of the inflow number or the outflow number is the reflux amount.

  Subsequently, the reflux amount estimation unit 14 adds the increase / decrease amount of the reflux amount in the analysis target year set in step S206 or S207 to the reflux amount in the previous year of the analysis target year, thereby calculating the reflux amount in the analysis target year. Obtain (S208). The obtained reflux amount is associated with a value indicating the year to be analyzed and stored in the auxiliary storage device 102 or the memory device 103, for example. The reflux amount in the initial reflux year is zero.

  Subsequently, the reflux amount estimation unit 14 advances the analysis target year by one year (S209). Subsequently, the reflux amount estimation unit 14 determines whether or not the analysis target year is larger than the latest year of possessed data (S210). The latest year of possessed data is a year indicating the time of the latest time-series data among the time-series data stored in the inflow / outflow number storage unit 121. When the analysis target year is less than or equal to the latest year of possessed data (NO in S210), Step S205 and subsequent steps are repeated. When the analysis target year is larger than the latest year of possessed data (YES in S210), the process of FIG. 6 ends.

  Note that the determination in step S210 may be performed by replacing the latest year of possessed data with the last year of the designated period designated for generation of the line model.

  Subsequently, a processing procedure executed by the influence amount estimation unit 15 will be described. FIG. 7 is a flowchart for explaining an example of a processing procedure executed by the influence amount estimation unit.

  In step S301, the influence amount estimation unit 15 acquires a line model of time-series data of the net increase of the company service. Subsequently, the influence amount estimation unit 15 extracts a change point of the broken line model (S302). For example, when the coordinate value of the change point is stored by the broken line model generation unit 13, the coordinate value is acquired. In addition, when there are a plurality of change points in the broken line model, Step S303 and subsequent steps are executed for each change point. Hereinafter, a change point that is a processing target is referred to as a “target change point”.

  Subsequently, the influence amount estimation unit 15 determines whether or not there is a change point in the same period (same month) as the target change point with respect to the broken line model of the selection data selected by the reflux amount estimation unit 14 in step S202. (S303).

  When there is a corresponding change point with respect to the polyline model of the selected data (YES in S303), the influence amount estimation unit 15 determines that the selected data (number of inflows or outflows) has been affected by the competitor service of the other company ( S304) The change amount of the inclination (number of persons / year) at the change point of the polygonal line model of the selection data (the inclination immediately after the change point−the inclination immediately before the change point) is calculated as the influence amount related to the selection data (S305). ).

  In other words, if the number of inflows or outflows changes when the net increase changes, the change is not based on recirculation, but is caused by the influence of other companies 'services (outflows or inflows related to other companies' services). Can be estimated. Therefore, it can be estimated that the change amount of the change is the influence amount of the service of other companies.

  On the other hand, when there is no corresponding change point with respect to the polyline model of the selected data (NO in S303), the influence amount estimation unit 15 is selected as selection data from the time-series data of the inflow number and the time-series data of the outflow number. It is determined that the time series data that has not been received (hereinafter referred to as “non-selected data”) has been influenced by the competitor's competing service (S306). For example, if the selected data is time-series data regarding the number of outflows, the time-series data regarding the number of inflows is non-selected data.

  Subsequently, the influence amount estimation unit 15 calculates the influence amount relating to the non-selected data from the speed change of the net increase and the speed change of the selected data based on the relational expression “net increase = inflow−outflow” (S307). . Specifically, when the outflow number side is selection data, the influence amount related to the inflow number can be obtained as follows based on the relational expression.

Influence amount related to the number of inflows = Speed change of the outflow number + Speed change of the net increase Here, the speed change of the outflow number and the speed change of the net increase are the speed change before and after the month related to the target change point (change in slope), respectively. ) And its value is positive or negative.

  Further, if the inflow number side is selection data, the influence amount relating to the outflow number can be obtained as follows.

Influence amount relating to outflow number = speed change in inflow number−speed change in net increase Note that step S307 is executed when the polygonal line model of the selection data does not have a change point at the time related to the target change point. That is, in step S307, when the speed change of the polyline model of the selected data at the time related to the target change point is 0, the non-selected data side is affected by the speed change of the positive or negative net increase. The speed change is assumed to be an influence amount on the non-selected data side. Therefore, as described above, when the inflow number side is the selection data, the influence amount related to the outflow number is calculated. If the outflow number side is selection data, the influence amount related to the inflow number is calculated in step S307.

  In addition, it can be estimated that the influence amount regarding the number of inflows or outflows of the company service calculated in step S305 or S307 is a speed change of the number of outflows from other company services or the number of inflows into other company services. In other words, it can be estimated that the influence amount related to the number of inflows of company services is a change in the number of outflows from other company services, and the influence amount related to the number of outflows of company services is a change in the number of inflows to other company services. .

  Subsequent to step S305 or S307, the influence amount estimation unit 15 obtains a line model of time series data of the number of subscribers of each company's service (S308). Subsequently, the influence amount estimation unit 15 determines whether the service of the other company related to the line model having a change point at the time (month) related to the target change point from among the line model of the time series data of the number of subscribers of the service of the other company. It is determined that the service is a service of another company that has influenced the influence amount in S305 or S307 (S309). For example, if the outflow rate of the company service is increasing and the net increase rate of the service of the other company is increasing, it can be determined that the customer outflowing from the company service has contracted (adopted) the service of the other company. The influence amount calculated in this case is a change in the number of outflows. In addition, when there are a plurality of applicable line model, it may be determined that a plurality of other company services related to the broken line model are the other company services that have influenced. In addition, it can be estimated that the service of another company that has influenced the number of inflows of the company service is, for example, the service of another company that directly or indirectly leaks into the company service. Indirect means a state in which another competitor service is interposed between the competitor service and the company service in the flow of the contractor from the company service to the company service. In addition, it can be estimated that the service of another company that has influenced the number of outflows of the company service is, for example, the service of another company directly or indirectly from the company service.

  As described above, according to the present embodiment, it is possible to grasp the movement of customers in the market. For example, when viewed from the business owner, it does not affect the net increase in the number of customers, but it is possible to estimate the amount of reflux that is a component that changes the number of inflows and outflows. In addition, based on the number of inflows or outflows that affect the net increase, it is possible to estimate the amount of influence received by the services of other companies on the services of other companies and the services of other companies that are affected by the influence amount.

  In the present embodiment, an example in which a service provider is a competitive element has been described. However, with respect to other competitive elements that can compete with other operators in the market, such as a business that sells products, Embodiments may be applied.

  In the present embodiment, the logarithmic time conversion unit 12 is an example of a conversion unit. The polygonal line model generation unit 13 is an example of a generation unit. The reflux amount estimation unit 14 is an example of a first estimation unit. The influence amount estimation unit 15 is an example of a second estimation unit.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to such specific embodiment, In the range of the summary of this invention described in the claim, various deformation | transformation・ Change is possible.

DESCRIPTION OF SYMBOLS 10 Estimation apparatus 11 Logarithmic origin setting part 12 Logarithmic time conversion part 13 Polygonal line model production | generation part 14 Reflux amount estimation part 15 Influence amount estimation part 100 Drive apparatus 101 Recording medium 102 Auxiliary storage apparatus 103 Memory apparatus 104 CPU
105 Interface Device 121 Inflow / Outflow Number Storage Unit 122 Contractor Number Storage Unit B Bus

Claims (7)

A conversion unit that converts time-series data regarding the number of inflows, outflows, and net increase of employers regarding the first competitive element into time-series data with a logarithmic time axis;
A generator that generates a line model approximating each converted time series data;
The amount of change in the slope of the line model related to the number of inflows or the line model related to the number of outflows at a time not corresponding to the change point of the line model related to the net increase does not affect the net increase, but the number of inflows Alternatively, a first estimation unit that estimates a reflux amount that is a component that changes the outflow number,
The estimation apparatus characterized by having. The first estimation unit is configured for a line model with a relatively small error with respect to the time-series data after conversion related to each line model among the line model related to the number of inflows or the line model related to the number of outflows. , Estimating the amount of change in the slope at a time not corresponding to the changing point of the polygonal line model related to the net increase as the reflux amount,
The estimation apparatus according to claim 1. A conversion unit that converts time-series data regarding the number of inflows, outflows, and net increase of employers regarding the first competitive element into time-series data with a logarithmic time axis;
A generator that generates a line model approximating each converted time series data;
If the second line model related to the number of inflows or the third line model related to the number of outflows has a change point at the time related to the change point of the first line model related to the net increase, the change point The amount of inflow from one of the one or more second competing elements competing with the first competing element for the amount of change in the slope of the second bent line model or the third broken line model at the time concerning Or a second estimation unit for estimating the influence amount relating to the number of outflows,
The estimation apparatus characterized by having. The second estimation unit, the second line model or the third line model, the line model of which the error is relatively smaller with the time-series data after conversion related to each line model, When there is no change point at the time related to the change point of the first broken line model, the amount of change in the slope of the first broken line model at the time is calculated as the error among the inflow number and the outflow number. Estimated as the influence amount from any of the second competitive factors with respect to the relatively larger one,
The estimation apparatus according to claim 3. The conversion unit further converts the time series data of the number of employers related to each of the second competitive elements into time series data with a logarithmic time axis,
The generation unit further generates a line model approximating the time series data after conversion relating to the number of employers of each of the second competitive elements,
The second estimation unit further relates to a line model having a change point at a time point related to a change point of the first line model among the line models related to the number of employers of the second competitive element. Presuming that the second competitive factor is an influence source related to the influence amount;
The estimation apparatus according to claim 3 or 4, characterized by the above. Computer
A conversion procedure for converting time-series data regarding the number of inflows, outflows, and net increase of employers regarding the first competitive element into time-series data with a logarithmic time axis;
Generation procedure to generate a line model approximating each time series data after conversion,
The amount of change in the slope of the line model related to the number of inflows or the line model related to the number of outflows at a time not corresponding to the change point of the line model related to the net increase does not affect the net increase, but the number of inflows Alternatively, a first estimation procedure for estimating the amount of reflux, which is a component that changes the outflow number,
The estimation method characterized by performing. Computer
A conversion procedure for converting time-series data regarding the number of inflows, outflows, and net increase of employers regarding the first competitive element into time-series data with a logarithmic time axis;
Generation procedure to generate a line model approximating each time series data after conversion,
If the second line model related to the number of inflows or the third line model related to the number of outflows has a change point at the time related to the change point of the first line model related to the net increase, the change point The amount of inflow from one of the one or more second competing elements competing with the first competing element for the amount of change in the slope of the second bent line model or the third broken line model at the time concerning Or a second estimation procedure for estimating the influence amount related to the number of outflows,
The estimation method characterized by performing.

Images