JP2008117012A - Information processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable prediction of a sales ratio etc., by agency for each commodity, and to enable calculation of a result reflecting performance data after the commodity has been sold. <P>SOLUTION: A predicted sales ratio calculating part 23 inputs, as a prior probability, a ratio represented by quantity sold of a specific commodity of each of a plurality of agencies set by a user in total quantity sold of the specific commodity sold by the plurality of agencies. A data reception part 21 receives performance data indicating a ratio represented by quantity sold of the specific commodity of each of the agencies in the total quantity sold of the specific commodity of the plurality of agencies as a result of the specific commodity sold actually in each of the plurality of agencies. A predicted sale ratio calculation part 23 calculates, as a predicted sales ratio, prediction on a ratio represented by the quantity sold of the specific commodity of each of the agencies, in the total quantity sold of the specific commodity of the plurality of agencies on the basis of the input prior probability and the performance data received with the data reception part 21. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information processing system that collects, aggregates, or stores information related to sales results of products for each of a plurality of agents.

  2. Description of the Related Art In recent years, an information processing system used in an insurance company has been developed, which has a feature that a plurality of agents having different sales characteristics are main sales channels, and products sold at each agent are severely modified.

  In such an information processing system, for example, a process of collecting, counting, or storing data (information) related to sales results (customer demographics, simplicity, or sales volume of products) for each of a plurality of agents is executed. .

  Insurers, as described above, the products sold at each agency are drastically modified, so the sales ratio of each product to be sold, etc., in order to consider the sales promotion policy for each agency. It is necessary to make various predictions about.

  An object of the present invention is to provide an information processing system capable of predicting a sales ratio or the like for each agent for each product and calculating a result reflecting actual data after the product is sold. .

  According to one aspect of the present invention, of the total number of sales of specific products sold by a plurality of agents, the ratio of the number of sales of the specific products for each of the plurality of agents set by the user Input means for inputting the value as a prior probability, and as a result of the fact that the specific product is actually sold in each of the plurality of agents, of the total number of sales of the specific product of the plurality of agents, the agent Receiving means for receiving performance data indicating a ratio of the number of sales of the specific product for each of the plurality of agents based on the prior probability input by the input means and the actual data received by the receiving means An information processing system comprising: a calculation unit that calculates, as a predicted sales ratio, a prediction of a ratio of the number of sales of the specific product for each agency out of the total sales of the specific product of the store It is provided.

  According to the present invention, it is possible to predict a sales ratio or the like for each agent for each product, and to calculate a result reflecting actual data after the product is sold.

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

[First Embodiment]
FIG. 1 is a block diagram showing a configuration of an information processing system 20 according to the first embodiment of the present invention. The information processing system 20 is a system used in an insurance company, for example. Hereinafter, the information processing system 20 will be described as the insurance company system 20. In addition, the insurance company system 20 can communicate with an information processing system (hereinafter referred to as an agency system) 10 used in an agency of an insurance company in which the insurance company system 20 is used. Note that there are a plurality of insurance company agents (agent systems used in). In FIG. 1, the description of the agency system other than the agency system 10 is omitted for convenience.

  The agency system 10 includes a sales record input unit 11, a sales record storage unit 12, and a data transmission unit 13.

  The sales performance input unit 11 inputs performance data indicating the sales performance of a specific product, for example, the product A, in response to an operation of a user who uses the agency system 10, for example, an employee of the agency. The result data input by the sales result input unit 11 is stored in the sales result storage unit 12. This performance data is input every day, for example.

  The data transmission unit 13 transmits the performance data stored in the sales performance storage unit 12 to the insurance company system 20.

  The insurance company system 20 includes a data receiving unit 21, a sales performance storage unit 22, a predicted sales rate calculation unit 23, a sales rate storage unit 24, and a predicted sales rate storage unit 25.

  The data receiving unit 21 receives the performance data transmitted by the data transmitting unit 13 of the agency system 10. Note that the performance data received here is performance data indicating the sales performance of the product A. In addition, the data receiving unit 21 receives performance data transmitted from a plurality of agency systems including the agency system 10. The record data received by the data receiver 21 is stored in the sales record storage unit 22.

  For example, when a newly developed product A is sold, the predicted sales ratio calculation unit 23 sells, for example, the product A for each of the agents out of the total number of products A sold by a plurality of agents. The ratio of the number (sales ratio) is input as the prior probability. This prior probability is a numerical value set (defined) by the user 100 as a target of the sales ratio of the product A for each agency, for example. For example, the user 100 operates the client terminal to set the prior probability. For example, the user 100 refers to the information stored in the sales ratio storage unit 24 and sets the ratio occupied by the number of sales of the product A for each agency.

  For example, the sales ratio storage unit 24 stores a sales ratio (information indicating) for each product sold in the past. This sales ratio indicates the ratio of the number of sales for each agency out of the total number of sales of the products corresponding to the sales ratio by a plurality of agencies. For example, when the product similar to the product A is the product a, the user 100 searches the sales rate storage unit 24 for the sales rate of the product a for reference when setting the prior probability. can do.

  The predicted sales ratio calculation unit 23 inputs numerical values A (X) to F (X) preset (defined) by the user 100, for example. These numerical values A (X) to F (X) are stored in a storage unit (not shown) of the insurance company system 20, for example. The numerical values A (X) to F (X) will be described later.

  The predicted sales ratio calculation unit 23, when the product A is sold at a plurality of agents based on the input prior probability and the result data stored in the sales result storage unit 22, the products of a plurality of agents Of the total sales number of A, the prediction of the ratio of the sales number of the product A for each of a plurality of agents is calculated as the predicted sales ratio. At this time, the predicted sales ratio calculation unit 23 calculates the predicted sales ratio according to the input numerical values A (X) to F (X). The predicted sales ratio calculation unit 23 stores various calculated values (described later) in the predicted sales ratio storage unit 25 when calculating the predicted sales ratio.

  In addition, for example, when the performance data is newly transmitted from a plurality of agency systems including the agency system 10, the forecast sales percentage calculation unit 23 and the various values stored in the forecast sales percentage storage unit 25 and the new data Based on the record data, for example, the predicted sales ratio is calculated (updated) every day. Moreover, the calculation result by the predicted sales ratio calculation unit 23 is presented (output) to the user 100, for example.

  Next, an example of performance data transmitted from the agency system 10 to the insurance company system 20 will be described with reference to FIG.

  As shown in FIG. 2, performance data is input to each of a plurality of agency systems including the agency system 10. As the actual data, for example, actual data including customer segment, simplicity, or sales volume of products a and b, for example, is input as a plurality of factors when products are sold at each agency.

  The customer group indicates, for example, the age group of customers who have purchased (contracted) products at the agency. For example, the customer segment includes a young segment, a middle-class segment, and an elderly segment. In addition, simplicity includes simplicity, ordinary, and complexity. In the example shown in FIG. 2, regarding the customer group, the agency A is 10% for the younger group, 60% for the middle-aged group, and 30% for the older group.

  The performance data input to each of the plurality of agency systems including the agency system 10 is transmitted to the insurance company system 20. The sales record storage unit 22 of the insurance company system 20 stores the customer segment, simplicity, etc. included in the record data. Further, the predicted sales ratio calculation unit 23 of the insurance company system 20 calculates the sales ratio (sales ratio results) for each product up to now according to the sales amount for each product included in the actual data. The calculated sales ratio results are stored, for example, in the sales result storage unit 22. The sales ratio performance indicates a ratio of the number of sales of the product a for each agency out of the number of sales of the product a at all the agencies, for example, as a result of the actual sale of the product a at each agency.

  The plurality of agency systems and the insurance company system 20 including the agency system 10 collect, aggregate, or store data (actual data) indicating the results of the actual sale of products at each of the plurality of agencies as described above. To do.

  Next, the processing procedure of the insurance company system 20 will be described with reference to FIG. Here, for example, it is assumed that a product A is developed as a new product of an insurance company. In addition, it is assumed that the product A is a product similar to the above-described product a, for example.

  First, for example, when the product A is sold by a plurality of agents, the user 100 sets a ratio (sales ratio) of the sales number of the product A for each agent in the total sales number of the product A. At this time, for example, the user 100 makes a setting with reference to information stored in the sales ratio storage unit 24. The ratio set by the user 100 is, for example, a sales ratio target (sales ratio target) of the product A for each agency.

  Here, since the product A is a product similar to the product a, the user 100 is set based on the sales record of the product a as shown in FIG. That is, it is assumed that the A agency is set to 40%, the B agency is set to 20%, the C agency is set to 20%, the D agency is set to 10%, and the E agency is set to 10%.

  The predicted sales ratio calculation unit 23 inputs the sales ratio target set by the user as a prior probability. Further, the predicted sales ratio calculation unit 23 inputs the numerical values A (X) to F (X) preset by the user 100 described above, for example.

  Here, it is assumed that the product A is actually sold at an agency (A to E). In the example shown in FIG. 3, the number of sales on the first day is 100 for the A agency, 100 for the B agency, 50 for the C agency, 50 for the D agency, and 200 for the E agency. As a result, the sales ratio performance of each agency on the first day is 20% for agency A, 20% for agency B, 10% for agency C, 10% for agency D, and 40% for agency E. . This data (actual data) is transmitted from the agency system of each agency to the insurance company system 20. Further, the sales ratio results are calculated by the predicted sales ratio calculation unit 23 based on the result data transmitted from a plurality of agency systems including the agency system 10.

  The predicted sales ratio calculation unit 23 uses, for example, Bayesian inference, for each agency based on the input prior probabilities, numerical values A (X) to F (X), and sales ratio results of each agency on the first day. Calculate the predicted sales ratio (predicted sales ratio after the first day). This predicted sales ratio is a prediction of the ratio of the number of sales of the product A for each of the plurality of agents out of the total sales of the product A of the plurality of agents when the product A is sold at each agency. It is shown.

  Here, it is assumed that the actual data of the next day (second day) is transmitted to the insurance company system 20. In this case, the predicted sales ratio calculation unit 23 calculates the second day of each agency based on the various values stored in the predicted sales ratio storage unit 25 and the sales ratio results on the day after the first day (second day). Calculate the expected sales ratio later. Similarly, the predicted sales ratio calculation unit 23 calculates the predicted sales ratio of each agency on and after the third day. That is, the predicted sales ratio calculation unit 23 updates the predicted sales ratio of each agency based on, for example, the daily sales ratio results.

  The predicted sales ratio calculated by the predicted sales ratio calculation unit 23 is presented to the user 100, for example.

  Next, with reference to the flowchart of FIG. 4, the processing procedure of the predicted sales ratio calculation unit 23 that calculates the predicted sales ratio using Bayesian inference will be described in detail.

  First, the predicted sales ratio calculation unit 23 inputs, for example, the sales ratio target (sales ratio target) of the product A for each agency set by the user 100 (step S1). The total sales ratio target for each agency is set to be 1. Hereinafter, for example, each agency is represented as an agency (X), and a sales ratio target for each agency (X) is represented as Y (X). That is, when the sales ratio target of the merchandise A of the agency A is 0.4, Y (A) = 0.4 is written. The following description will be made in the same manner for each numerical value.

  Next, the predicted sales ratio calculation unit 23 inputs the numerical values A (X) to F (X) set by the user 100 described above (step S2).

  Here, numerical values A (X) to F (X) will be described. The numerical value A (X) is a probability initial value at which the agency (X) achieves the sales ratio target Y (X) set by the user 100, for example, and is a numerical value arbitrarily set by the user. The numerical value B (X) is a probability initial value at which the agency (X) cannot achieve the sales ratio target Y (X) set by the user 100, for example. The numerical value B (X) is a numerical value arbitrarily set by the user, like the numerical value A (X), and is 1-A (X). In the following description, it is assumed that A (X) = 0.5 and B (X) = 0.5.

  In the numerical values C (X) to F (X), probabilities are arbitrarily set for assumed combinations. The numerical value C (X) is a probability that the sales ratio target Y (X) is finally achieved, and the sales ratio performance has achieved the sales ratio target Y (X). Here, C (X) = Y (X).

  The numerical value D (X) is a case where the sales ratio target Y (X) is finally achieved, and the sales ratio performance does not achieve the sales ratio target Y (X). Here, C (X) = 1−Y (X).

  The numerical value E (X) is a probability when the sales ratio target Y (X) is not finally achieved, and the sales ratio performance has achieved the sales ratio target Y (X). Here, it is assumed that E (X) = 0.1 (default value: arbitrary) is set.

  The numerical value F (X) is a probability in the case where the sales ratio target Y (X) is not finally achieved, and the sales ratio performance does not achieve the sales ratio target Y (X). Here, it is assumed that F (X) = 1−E (X) = 0.9 (default value: arbitrary) is set.

  Note that the above-described numerical values A (X) to F (X) are set in advance, and for example, the user 100 can change each numerical value.

  Here, it is assumed that the sales ratio target Y (X) of the merchandise A of the agency (X) is 0.4, so that C (X) = 0.4 and D (X) = 0,6. It will be explained as a thing.

  Next, the predicted sales ratio calculation unit 23 determines the sales ratio (ratio) according to the number of sales of the product A of, for example, the agency (X) included in the actual data stored in the sales result storage unit 22 (today). ) Actual I (X) is calculated (step S3).

  The predicted sales ratio calculation unit 23 compares, for example, the sales ratio performance I (X) (of the day) and the sales ratio target Y (X) of the agency (X) (step S4).

  Here, when the sales ratio performance I (X) is larger than the sales ratio target Y (X) (= C (X)) (YES in step S5), the predicted sales ratio calculation unit 23 calculates A (X) × C ( X): B (X) × E (X) is calculated. In this case, if A (X) = 0.5, C (X) = 0.4, B (X) = 0.5, and E (X) = 0.1 as described above, 0.2 : 0.05. The predicted sales ratio calculation unit 23 converts the calculated 0.2: 0.05 so that the whole becomes 1, and replaces it with new A (X) and B (X) (step S6). That is, 0.2: 0.05 = 0.8: 0.2 = A (X): B (X).

  On the other hand, when the sales ratio performance I (X) is smaller than the sales ratio target Y (X) (= C (X)) (NO in step S5), the predicted sales ratio calculation unit 23 calculates A (X) × D (X ): B (X) × F (X) is calculated. In this case, if A (X) = 0.5, D (X) = 0.6, B (X) = 0.5, and F (X) = 0.9 as described above, 0.3 : 0.45. The predicted sales ratio calculation unit 23 converts the calculated 0.3: 0.45 so that the whole becomes 1, and replaces it with new A (X) and B (X) (step S7). That is, 0.3: 0.45 = 0.4: 0.6 = A (X): B (X).

  The predicted sales ratio calculation unit 23 stores A (X) and B (X) calculated in step S6 or S7 in the predicted sales ratio storage unit 25.

  The agents (X) are ranked in the order of higher rank of A (X) calculated in step S6 or S7 (step S8). Here, since A (X) is the probability (excellence) that the agency (X) achieves the sales ratio target Y (X), the agency (X) having the higher value of A (X) Can be evaluated as an excellent agency.

  Here, for example, as described above, when A (X) is 0.8, it is assumed that C (X) (= Y (X)) is achieved, and 2 × C (X ) × A (X) = G (X) is assumed as the target expected value (step S9). In this case, for example, when C (X) = 0.4 and A (X) = 0.8 are substituted, G (X) = 0.64.

  Here, for example, it is assumed that the agency (X) includes an agency (A), an agency (B), and an agency (C). In this case, for example, it is assumed that G (A) = 0.64, G (B) = 0.264, and G (C) = 0.44. In this case, the predicted sales ratio calculation unit 23 converts (converts) the target expected value (G (X)) for each agency (X) so as to be 1 as a whole, and converts the converted numerical value to H (X ). In this case, H (A) = 0.48, H (B) = 0.2, and H (C) = 0.32. The predicted sales ratio calculation unit 23 calculates this H (X) as the predicted sales ratio (step S10).

  Agents (X), H (X), and Y (X) are presented in ascending order of the predicted sales ratio H (X) calculated by the predicted sales ratio calculation unit 23 (step S11).

  For example, when actual data is newly stored in the sales result storage unit 22 on the next day (second day) (when transmitted from a plurality of agency systems including the agency system 10 to the insurance company system 20). , A (X) and B (X) are replaced with A (X) and B (X) stored in the predicted sales ratio storage unit 25, and the process returns to step S3 and is repeated.

  FIG. 5 is a diagram illustrating an example of a calculation result up to the predicted sales ratio after the third day of each agency by the predicted sales ratio calculation unit 23.

  As shown in the example of FIG. 5, for example, in the agency A, from the prior probability Y (A) 40% and the sales ratio (ratio) result I (A) 20% on the first day, the predicted sales ratio H (A) after the first day 29% is calculated. Further, the predicted sales ratio H (A) 22% after the second day is calculated from the sales ratio performance I (A) 10% on the second day. Similarly, the predicted sales ratio H (A) 41% after the third day is calculated from the sales ratio performance I (A) 40% on the third day.

  In addition, in the E agency, for example, even if the sales ratio performance I (E) on the first day is abnormally high (40%) (abnormal value), the predicted sales ratio H after the first day using the prior probability Y (E). Since (E) is calculated, the influence of the abnormal value can be reduced.

  FIG. 6 is a diagram illustrating an example of a case where the calculation result as illustrated in FIG. 5 is displayed on a display screen (display) of a client terminal operated by the user 100, for example. In addition, the structure displayed on the display part (display) with which the insurance company system 20 is provided may be sufficient.

  As shown in FIG. 6, for example, on the display screen of the client terminal operated by the user 100, the agency (X), the predicted sales ratio H (X), and the advance sales according to the ranking of the predicted sales ratio H (X). Probability Y (X) is shown. Here, it is displayed based on the predicted sales ratio H (X) after the third day of each agency shown in FIG.

  In the example shown in FIG. 6, the predicted sales ratio H (X) of the agency (X) whose predicted sales ratio H (X) is higher than the prior probability Y (X) and the predicted sales ratio based on the ranking based on the prior probability Y (X). The ranking of the agency (X) with the highest ranking by H (X), the forecast sales ratio H (X) and the prior probability of the agency (X) whose forecast sales ratio H (X) is lower than the prior probability Y (X) The ranking of the agency (X) lower than the ranking based on Y (X) is displayed separately. Further, for example, when the prior probability Y (X) and the predicted sales ratio H (X) are the same, or when the ranking based on the prior probability Y (X) and the ranking based on the predicted sales ratio H (X) are the same, distinction is made. It is displayed as it is.

  As described above, in the present embodiment, for example, Bayesian inference is used. Therefore, for example, the predicted sales ratio (predicted sales ratio after the first day) for each agent of the product A can be calculated using the prior probability. In addition, by reflecting the subsequent sales ratio performance to the calculated predicted sales ratio, it is possible to calculate a result (predicted sales ratio) with correction each time. For example, the insurance company (user 100) can select an excellent sales prospective agent for planning or correcting the sales promotion activity policy of the product A based on the calculated result.

  Further, in the present embodiment, for example, even an insurance product having a characteristic that the life cycle of the product is short, the result can be fed back with the prior probability and a small amount of actual data (sales ratio actual). Further, by using the prior probability referring to the past performance data stored in the sales ratio storage unit 24, for example, the influence on the calculation by the abnormal value can be reduced.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. FIG. 7 is a block diagram showing the configuration of the insurance company system 30 according to the present embodiment. 1 are denoted by the same reference numerals, and detailed description thereof is omitted. Here, parts different from FIG. 1 will be mainly described. The same applies to the following embodiments.

  The insurance company system 30 includes a Bayesian network 31, a predicted sales ratio calculation unit 32, an effective sales factor calculation unit 33, and an agency characteristic storage unit 34.

  In the Bayesian network 31, for example, when the product A is sold, for each of a plurality of factors (for example, customer segment and simplicity), an influence degree indicating the influence of the factor on the sales of the product A is set as a prior probability. (Input). This prior probability is a numerical value set by the user 100. Details of the Bayesian network 31 will be described later.

  The predicted sales ratio calculation unit 32 calculates, for example, a predicted sales ratio (first predicted sales ratio) for each agent of the product A based on the Bayesian network 31. This predicted sales ratio indicates a prediction of the ratio of the total sales number of the product A to the sales number of the product A for each agency. Hereinafter, this predicted sales ratio will be described as a prior probability predicted sales ratio. This prior probability predicted sales ratio is presented to the user 100, for example.

  The predicted sales ratio calculation unit 32 calculates a predicted sales ratio (second predicted sales ratio) for each agency based on the calculated prior probability predicted sales ratio and the actual data stored in the sales result storage unit 22. To do. The predicted sales ratio calculation unit 32 stores the calculated various values (A (X) and B (X) in the first embodiment described above) in the predicted sales ratio storage unit 25 when calculating the predicted sales ratio. To do.

  Moreover, when the performance data is newly transmitted from a plurality of agency systems including the agency system 10, for example, the predicted sales ratio calculation unit 23 stores various values (A (X ) And B (X)) and the new result data, for example, the predicted sales ratio is calculated (updated) every day.

  Based on the predicted sales ratio calculated by the predicted sales ratio calculation unit 32 and the Bayesian network 31, the effective sales factor calculation unit 33 back-calculates the degree of influence indicating the influence on the sales of the product A for each of a plurality of factors, for example. . The effective sales factor calculation unit 33 extracts, for example, a factor (effective sales factor) that is effective when the product A is sold, according to the degree of influence calculated backward. The effective sales factor extracted by the effective sales factor calculation unit 33 is presented to the user 100, for example.

  For each factor of sales of products sold in the past, the agent characteristic storage unit 34 sells products sold according to the factor of each agent out of the total number of sales sold according to the factor. Stores ratio information indicating the ratio of numbers. When the user 100 predicts the degree of influence for each factor as described above, for example, the user 100 searches past data (ratio information) stored in the agency characteristic storage unit 34 and refers to the searched data. be able to.

  Next, the processing procedure of the insurance company system 30 will be described with reference to FIG. The past performance data will be described with reference to FIG.

  Here, for example, it is assumed that a product A is developed as a new product of an insurance company. Further, it is assumed that the product A is a product similar to the product a shown in FIG. For convenience, the description will be made assuming that the plurality of agents are the A agent and the B agent.

  First, for example, when the product A is sold, the user 100 sets an influence degree indicating an influence on the sale for each of a plurality of factors.

  For example, the degree of influence set by the user 100 is set (input) as a prior probability in the Bayesian network 31.

  Here, FIG. 9 is a diagram for explaining the Bayesian network 31. As shown in FIG. 9, by setting the prior probabilities (here, X and Y) in the Bayesian network 31, for example, out of the total number of sales when the product A is sold, for each agency (A agency, Prediction of the ratio of the number of sales of (B agency) (prior probability predicted sales ratio) is calculated.

  Note that the ratios of simplicity and customer demographics for each agency shown in FIG. 9 are, for example, the ratios of “simplified” simplicity and “young” demographics shown in FIG. ) In advance. Further, in the “young” group of customers shown in FIG. 2, the A agency is 10% and the B agency is 80%, but for convenience, the A agency is 10% (0.1). Is described as 90% (0.9).

  In other words, the prior probability predicted sales ratio is calculated based on the prior probability and ratio information set in the Bayesian network 31.

  Based on the Bayesian network 31 shown in FIG. 9, when the prior probability predicted sales ratio of the agency A and the agency B is calculated, the prior probability predicted sales ratio of the agency A is calculated by 0.8X + 0.1Y. . Further, the prior probability predicted sales ratio of the agency B is calculated by 0.2X + 0.9Y.

  Here, for example, it is assumed that the user 100 predicts that the product A has a high degree of dependence (influence) on simplicity. Accordingly, for example, it is assumed that X = 0.7 and Y = 0.3 are set in the Bayesian network 31 as prior probabilities. In this case, the prior probability predicted sales ratio of the agency A is 59%. Similarly, the prior probability predicted sales ratio of the agency B is 41%.

  Here, FIG. 10 is a diagram illustrating an example when the prior probability predicted sales ratio calculated based on the Bayesian network 31 is displayed on the display screen of the client terminal operated by the user 100, for example.

  When X = 0.7 and Y = 0.3 are set (input) for the Bayesian network 31, for example, the display screen displayed (presented) to the user 100 includes the A agent and the B agent. 59% and 41% are displayed as the prior probability predicted sales ratios.

  A description will be given with reference to FIG. 8 again. As shown in FIG. 8, when product A is sold at A and B agents, the number of products A sold at the A agent is 40 on the first day, and the number of products A sold at B agent is on the first day. Suppose that it was 160. Accordingly, it is assumed that the sales ratio performance on the first day of the agency A is 20% and the sales ratio performance on the first day of the agency B is 80%.

  The predicted sales ratio calculation unit 32 calculates the predicted sales ratio of the A and B agents based on the prior probability predicted sales ratio and the sales ratio performance of the A and B agents, for example, by using Bayesian inference. The method for calculating the predicted sales ratio is the same as that in the case where the prior probability in the first embodiment described above is replaced with the prior probability predicted sales ratio in the present embodiment, and the description thereof is omitted. That is, also in this embodiment, as in the first embodiment described above, for example, numerical values A (X) to F (X) are set in advance.

  Here, FIG. 11 is a diagram illustrating an example of a calculation result of the predicted sales ratio several days after the sales start of the products A of the A and B agents by the predicted sales ratio calculation unit 32.

  As shown in FIG. 11, the predicted sales ratio several days after the sales start of the merchandise A of the agency A is 24%. In addition, the predicted sales ratio several days after the sales start of the merchandise A of the agency B is 76%.

  Referring again to FIG. 8, the effective sales factor calculation unit 33 shows the influence on the sales of the product A for each of a plurality of factors based on the predicted sales rate calculated by the predicted sales rate calculation unit 32 and the Bayesian network 31. Back-calculate the impact.

  As shown in FIG. 11, for example, when the predicted sales ratio several days after the sales start of the merchandise A of the agency A is 24%, and the predicted sales ratio several days after the sales start of the merchandise A of the agency B is 76% Is assumed. In this case, when the predicted sales ratio of the agency A is applied to the Bayesian network 31, 0.24 (24%) = 0.8X + 0.1Y. Further, when the predicted sales ratio of the agency B is applied to the Bayesian network 31, 0.76 (76%) = 0.2X + 0.9Y. From these, X = 0.2 and Y = 0.8 are calculated backward. Thereby, the effective sales factor calculation part 33 extracts the factor (effective sales factor) which becomes effective, for example at the time of sale of the goods A.

  The effective sales factor extracted by the effective sales factor calculating unit 33 is presented to the user 100, for example.

  FIG. 12 is a diagram illustrating an example when the effective sales factor extracted by the effective sales factor calculating unit 33 is displayed on the display screen of the client terminal operated by the user 100, for example.

  As shown in FIG. 12, on the display screen displayed for the user 100, for example, an effective sales factor is displayed separately from other factors. For example, when the back calculation result as described above is calculated, customer segments are displayed separately as shown in FIG. In addition, a guidance display shows that the effective sales factor calculation unit 33 can know from the result of the reverse calculation. In the example of FIG. 12, for example, “(Product A) is strongly customer-oriented and needs to strengthen sales to younger generations” is displayed.

  For example, by confirming the display screen shown in FIG. 12, the user 100 is more likely to rely on the customer segment for the product A than the simplicity of the initial forecast, and the sales forecast at the B agency that is strong in the younger segment. You can recognize that youth is well-oriented.

  As described above, in the present embodiment, for example, Bayesian inference is used. Therefore, for example, an effective sales factor of the product A can be extracted by using the prior probability (expected value) and the Bayesian network. Further, by reflecting the actual data after the sale of the product A, it is possible to calculate the result of the correction each time. For example, the insurance company (user 100) can plan or modify the sales promotion activity policy for the product A based on this result.

  Further, in the present embodiment, for example, even an insurance product having a characteristic that the life cycle of the product is short, the result can be fed back with a prior probability and a small amount of actual data. Moreover, the influence on the calculation by an abnormal value can be relieved by using the prior probability which referred the past performance data. In addition, since a Bayesian network leading to the result is defined, it is possible to analyze the progress from the result.

[Third Embodiment]
Next, a third embodiment of the present invention will be described. FIG. 13 is a block diagram showing a configuration of the insurance company system 50 according to the present embodiment. The insurance company system 50 can communicate with the agency system 40. Note that although there are a plurality of agents of the insurance company, in FIG. 13, the description of the agent system other than the agent system 40 is omitted for convenience.

  The agency system 40 includes a sales record input unit 41, a sales record storage unit 42, and a data transmission unit 43.

  The sales record input unit 41 inputs record data indicating the sales record of a plurality of specific products, for example, the products A and B, according to the operation of a user who uses the agent system 40, for example, an employee of the agent, for example. . The result data input by the sales result input unit 41 is stored in the sales result storage unit 42. This performance data is input every day, for example.

  The data transmission unit 43 transmits the performance data stored in the sales performance storage unit 42 to the insurance company system 50.

  The insurance company system 20 includes a data receiving unit 51, a sales performance storage unit 52, a Bayesian network 53, and a predicted sales ratio calculation unit 54.

  The data receiving unit 51 receives the result data transmitted by the data transmitting unit 43 of the agency system 40. The performance data received here is performance data indicating the sales performance of the products A and B. In addition, the data receiving unit 51 receives performance data transmitted from an agent system used by a plurality of agents. The record data received by the data receiver 51 is stored in the sales record storage unit 52.

  In the Bayesian network 53, for example, when products A and B are sold, for each of a plurality of factors (for example, customer segment and simplicity), the degree of influence indicating the influence of the factors on the sales of the products A and B is shown. It is set (input) as a prior probability (first prior probability). Further, for example, when the products A and B are sold, the ratio of the sales number of the products A or B for each agency, for example, out of the total sales number of the products A and B sold by a plurality of agents is It is set as a prior probability (second prior probability). This prior probability (the first prior probability and the second prior probability) is a numerical value set by the user. Details of the Bayesian network 53 will be described later.

  Based on the Bayesian network 53, the predicted sales ratio calculation unit 54 calculates, for example, a predicted sales ratio (second predicted sales ratio) of the product A or B. The predicted sales ratio indicates, for example, a prediction of a ratio of the sales number of the products A and B to the sales number of the product A or B. In addition, the predicted sales ratio is presented to the user 100, for example. Further, based on the Bayesian network 53, for example, a predicted sales ratio (first predicted sales ratio) for each agent of the products A and B is calculated as in the second embodiment described above. The first predicted sales ratio corresponds to the prior probability predicted sales ratio in the second embodiment described above.

  The predicted sales ratio calculation unit 54 calculates the number of sales of the products A and B for each agency indicated by the calculated predicted sales ratio (first predicted sales ratio) and the actual data stored in the sales result storage unit 52. Based on the ratio, a predicted sales ratio (third predicted sales ratio) for each agency is calculated. The predicted sales ratio calculation unit 54 stores the calculated various values (A (X) and B (X)) in the predicted sales ratio storage unit 25 when calculating the predicted sales ratio.

  Further, the predicted sales ratio calculation unit 23, for example, when new performance data is transmitted from a plurality of agency systems including the agency system 40, various values (A (X ) And B (X)) and the new result data, for example, the predicted sales ratio is calculated (updated) every day.

  Further, the predicted sales ratio calculation unit 54 calculates the predicted sales ratio (fourth predicted sales ratio) of the products A and B based on the degree of influence calculated by the effective sales factor calculation unit 33. The predicted sales ratio calculated by the predicted sales ratio calculation unit 54 is presented to the user 100, for example.

  Next, the processing procedure of the insurance company system 50 will be described with reference to FIG. The past performance data will be described with reference to FIG.

  Here, for example, products A and B are developed as new products of an insurance company. Further, it is assumed that the product A is similar to the product a shown in FIG. 2, for example, and the product B is a product similar to the product b shown in FIG. For convenience, the description will be made assuming that the plurality of agents are the A agent and the B agent.

  First, for example, when the products A and B are sold, the user 100 sets an influence degree indicating the influence on the sales of the products A and B for each of a plurality of factors.

  In addition, when the products A and B are sold by the A and B agents, the user 100 occupies the ratio of the sales number of the products A or B for each agent in the total sales number of the products A and B (sales Set the ratio. At this time, for example, the user 100 makes a setting with reference to information stored in the sales ratio storage unit 24.

  In the Bayesian network 31, for example, an influence degree and a ratio set by the user 100 are set (input) as a prior probability.

  Here, FIG. 15 is a diagram for explaining the Bayesian network 53. As shown in FIG. 15, by setting prior probabilities (here, X, Y, Z, W, S, T, U and V) in the Bayesian network 53, for example, when products A and B are sold. A prediction (second predicted sales ratio) of the ratio of the sales volume of the product A or B in the total sales volume is calculated. At this time, as in the second embodiment described above, for example, the ratio of the number of sales for each agency (A agency, B agency) to the total sales quantity when the products A and B are sold. Is calculated (first predicted sales ratio).

  Note that the simplicity, which is the plurality of factors shown in FIG. 15, and the ratio of the customer segment to each agency are the same as those in the second embodiment described above, and thus the description thereof is omitted.

  When calculating the second predicted sales ratio of the products A and B based on the Bayesian network 53 illustrated in FIG. 15, the prior probability predicted sales ratio of the product A is 0.8XS + 0.2XU + 0.1YS + 0.9YU + 0.8ZS + 0. Calculated by 2ZU + 0.1WS + 0.9WU. Further, the second predicted sales ratio of the product B is calculated by 0.8XT + 0.2XV + 0.1YT + 0.9YV + 0.8ZT + 0.2ZV + 0.1WT + 0.9WV.

  Here, for example, the user 100 predicts that the product A has a high degree of dependence (influence) on simplicity, and for example, X = 0.7 and Y = 0.3 are set in the Bayesian network 53 as prior probabilities. And For example, it is assumed that the user 100 predicts that the product B does not have a specific dependency, and sets Z = 0.5 and W = 0.5 in the Bayesian network 53 as prior probabilities, for example. Further, since the product A is similar to the product a and the product B is similar to the product b, the user 100 predicts from the sales amount of the products a and b shown in FIG. , T = 0.3, U = 0.5, and V = 0.5 are set in the Bayesian network 53. In this case, the second predicted sales ratio of the product A is 60%. Similarly, the second predicted sales ratio of the product B is 40%. This second predicted sales ratio is presented to the user 100, for example.

  Here, FIG. 16 is a diagram illustrating an example when the second predicted sales ratio calculated based on the Bayesian network 53 is displayed on the display screen of the client terminal operated by the user 100, for example. .

  For example, as described above for the Bayesian network 31, X = 0.7, Y = 0.3, Z = 0.5, W = 0.5, S = 0.7, T = 0.3, U = 0. .5 and V = 0.5 are set (input), for example, on the display screen displayed (presented) to the user 100, the second predicted sales ratio of the products A and B is 60% and 40% is displayed.

  A description will be given with reference to FIG. 14 again. As shown in FIG. 813, when the products A and B are sold at the A and B agents, the number of sales on the first day of the product A at the agent A is 100, and the number of sales of the product B on the first day is 70. Shall be. Further, it is assumed that the number of sales of the product A on the first day at the agency B is 100 and the number of sales of the product B on the first day is 50. In this case, the sales ratio performance on the first day of the agency A is about 55%, and the sales ratio performance on the first day of the agency B is about 45%.

  Hereinafter, in the same manner as in the first and second embodiments described above, the predicted sales ratio calculation unit 54 uses the Bayesian inference, for example, to calculate the third predicted sales ratio and the sales ratio results from the first Calculate the predicted sales ratio. That is, also in the present embodiment, for example, numerical values A (X) to F (X) are set in advance as in the first and second embodiments described above. Further, as in the second embodiment described above, the effective sales factor calculation unit 33 calculates the degree of influence indicating the influence on the sales of the products A and B for each of a plurality of factors. As a result, the values of X, Y, Z, and W are, for example, X = 0.3, Y = 0.7, Z = 0. 6. It is assumed that W is updated to 0.4. In addition, the values of S, T, U, and V are updated to, for example, S = 0.2, T = 0.8, U = 0.5, and V = 0.5 depending on the sales performance ratio of the A and B agents. Shall be.

  When each value is updated as described above, the predicted sales ratio calculation unit 54 calculates the predicted sales ratio (fourth predicted sales ratio) of the products A and B based on the value. For example, when updated to the above-described value, the predicted sales ratio of the product A is 38%, and the predicted sales ratio of the product B is 62%.

  The predicted sales ratio calculated by the predicted sales ratio calculation unit 54 is presented to the user 100, for example.

  FIG. 17 is a diagram illustrating an example when the predicted sales ratio calculated by the predicted sales ratio calculation unit 54 is displayed on the display screen of the client terminal operated by the user 100, for example.

  As shown in FIG. 17, the display screen displayed (presented) to the user 100 includes, for example, a product with a high fourth predicted sales ratio calculated by the predicted sales ratio calculation unit 54 among the products A and B. However, it is displayed separately from other products as a better-selling product.

  For example, by checking the display screen shown in FIG. 17, the user 100 can recognize that the product B is more popular than the product A. Therefore, the user 100 can plan and modify the sales promotion activity policy associated therewith.

  As described above, in the present embodiment, for example, Bayesian inference is used, and therefore, for example, the best selling item can be selected by using the prior probability (expected value) and the Bayesian network. Further, by reflecting the actual data after sales of the products A and B, it is possible to calculate the result of correction each time. For example, the insurance company (user 100) can plan or revise sales promotion activity policies for the products A and B based on the result.

  Further, in the present embodiment, for example, even an insurance product having a characteristic that the life cycle of the product is short, the result can be fed back with a prior probability and a small amount of actual data. Moreover, the influence on the calculation by an abnormal value can be relieved by using the prior probability which referred the past performance data. In addition, since a Bayesian network leading to the result is defined, it is possible to analyze the progress from the result.

  Note that the present invention is not limited to the above-described embodiments as they are, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Further, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in each embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

1 is a block diagram showing a configuration of an information processing system 20 according to a first embodiment of the present invention. The figure which shows an example of the performance data transmitted to the insurance company system 20 from the agency system 10. The flowchart which shows the process sequence of the insurance company system. The flowchart which shows the process sequence of the prediction sales ratio calculation part 23. FIG. The figure which shows an example of the calculation result by the prediction sales ratio calculation part 23. FIG. The figure which shows an example at the time of displaying the calculation result by the prediction sales ratio calculation part 23 with respect to the user 100. FIG. The block diagram which shows the structure of the insurance company system 30 which concerns on the 2nd Embodiment of this invention. The flowchart which shows the process sequence of the insurance company system. The figure for demonstrating the Bayesian network 31. FIG. The figure which shows an example in case the prior probability prediction sales ratio calculated based on the Bayesian network 31 is displayed with respect to the user 100, for example. The figure which shows an example of the calculation result by the prediction sales ratio calculation part 32. FIG. The figure which shows an example when the result calculated by the effective sales factor calculation part 33 is displayed with respect to the user 100, for example. The block diagram which shows the structure of the insurance company system 50 which concerns on the 3rd Embodiment of this invention. The flowchart which shows the process sequence of the insurance company system. The figure for demonstrating the Bayesian network 53. FIG. The figure which shows an example in case the prior probability prediction sales ratio calculated based on the Bayesian network 53 is displayed with respect to the user 100, for example. The figure which shows an example at the time of the prediction sales ratio calculated by the prediction sales ratio calculation part 54 being displayed with respect to the user 100, for example.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10,40 ... Agency system, 11, 41 ... Sales result input part, 12, 42 ... Sales result storage part, 13, 43 ... Data transmission part, 20, 30, 50 ... Insurance company system, 21, 51 ... Data reception , 22, 52 ... sales result storage unit, 23 ... predicted sales rate calculation unit, 24 ... sales rate storage unit, 25 ... predicted sales rate storage unit, 31 ... Bayesian network, 32 ... predicted sales rate calculation unit, 33 ... effective Sales factor calculation unit 34... Agent characteristic storage unit 53. Bayesian network 54.

Claims (6)

An input means for inputting, as a prior probability, a ratio of the number of sales of the specific product for each of the plurality of agents set by the user out of the total number of sales of the specific product sold by the plurality of agents;
As a result of the actual sale of the specific product in each of the plurality of agents, the total number of sales of the specific product of the plurality of agents is the number of sales of the specific product for each of the agents. A receiving means for receiving performance data indicating the proportion of occupancy;
Based on the prior probability input by the input means and the actual data received by the receiving means, out of the total number of sales of the specific goods of the plurality of agents, the specific goods for each of the agents An information processing system comprising: a calculation unit that calculates a prediction of a ratio occupied by the number of sales as a predicted sales ratio.   The information processing system according to claim 1, further comprising a presentation unit that presents the predicted sales ratio calculated by the calculation unit to the user. For each of the multiple factors when a product is sold, the ratio of the number of products sold according to the factor of each agency to the total number of products sold according to the factor. Storage means for storing the percentage information shown;
For each of a plurality of factors when a specific product is sold, an input means for inputting, as a prior probability, an influence indicating the influence of the factor set by the user on the sales of the specific product;
Based on the ratio information stored in the storage means and the prior probability input by the input means, out of the total sales number of the specific product sold by the plurality of distributors, A first calculation means for calculating a prediction of a ratio of the number of sales of a specific product as a first predicted sales ratio;
As a result of actual sales of the specific product at each of the plurality of agents, out of the total number of sales of the specific product of the plurality of agents, the number of sales of the specific product for each of the agents A receiving means for receiving performance data indicating a ratio;
Based on the first predicted sales ratio calculated by the first calculation means and the actual data received by the receiving means, the agency out of the total number of sales of the specific product of the plurality of agencies A second calculating means for calculating a prediction of a ratio of the number of sales of the specific product for each as a second predicted sales ratio;
Based on the second predicted sales ratio calculated by the second calculating means, the degree of influence indicating the influence of each of a plurality of factors when the specific product is sold on the sales of the specific product A third calculating means for calculating;
An information processing system comprising: an extracting unit that extracts a factor that is effective in selling the specific product among the plurality of factors according to the degree of influence calculated by the third calculating unit. system. First presenting means for presenting to the user the first predicted sales ratio calculated by the first calculating means;
The information processing system according to claim 3, further comprising: a second presenting unit that presents the factor extracted by the extracting unit to the user. For each of the multiple factors when a product is sold, the ratio of the number of products sold according to the factor of each agency to the total number of products sold according to the factor. Storage means for storing the percentage information shown;
For each of a plurality of factors when a plurality of specific products are sold, an influence degree indicating the influence of the factor set by the user on the sales of each of the plurality of specific products is input as a first prior probability First input means to:
Of the total number of sales of a plurality of specific products sold by a plurality of agents, the ratio of the number of sales of each of the plurality of specific products for each of the plurality of agents set by the user is the second A second input means for inputting as a prior probability;
Based on the ratio information stored in the storage unit and the first prior probability input by the first input unit, the total number of sales of the plurality of specific products sold by the plurality of agents Of these, a first calculation means for calculating a prediction of a ratio of the number of sales of each of the plurality of specific products for each agency as a first predicted sales ratio;
The plurality of identifications based on the ratio information stored in the storage means, the first prior probability input by the first input means and the second prior probability input by the second input means A second calculation means for calculating a prediction of a ratio of the sales of each of the plurality of specific products as a second predicted sales ratio among the total sales of the products of
As a result of actually selling each of the plurality of specific products at each of the plurality of agents, the ratio of the number of sales of each of the plurality of specific products for each of the agents and all of the plurality of agents Receiving means for receiving performance data indicating the proportion of the number of sales of each of the plurality of specific products in
Based on the first predicted sales ratio calculated by the first calculation means and the ratio of the number of sales of each of the plurality of specific products for each agency indicated by the performance data received by the receiving means. And calculating a prediction of a proportion of the total number of sales of the plurality of specific products of the plurality of agents by the number of sales of the plurality of specific products for each of the agents as a third predicted sales ratio. 3 calculation means;
Based on the first predicted sales ratio calculated by the first calculation means, each of the plurality of factors when each of the plurality of specific products is sold is the sales of each of the plurality of specific products. Fourth calculating means for calculating the degree of influence indicating the influence on
Proportion occupied by the number of sales of each of the plurality of specific products indicated by the ratio information stored in the storage means, the degree of influence calculated by the fourth calculation means, and the performance data received by the receiving means And a fifth calculating means for calculating, as a fourth predicted sales ratio, a prediction of a ratio occupied by each sales number of the plurality of specific products out of the total sales number of the plurality of specific products. An information processing system comprising: First presenting means for presenting to the user the second predicted sales ratio calculated by the second calculating means;
The information processing system according to claim 5, further comprising: a second presentation unit that presents the fourth predicted sales ratio calculated by the fifth calculation unit to the user.

Images