Title of the invention: Predictive business support device and predictive business support method
Technical field
[0001]
　The present invention relates to a predictive business support device and a predictive business support method.
Background technology
[0002]
　In financial institutions, there is a so-called ALM (asset liability management) business. This ALM business predicts certain events (eg, official discount rate, stock price, etc.) based on the composition of assets and liabilities such as deposit amount and loan lending interest rate at financial institutions, and based on the prediction results, various risks It aims to minimize and maximize profits. Therefore, the person in charge of ALM business makes future forecasts while referring to various economic indicators and analyst reports.
[0003]
　As a conventional technique related to such future forecasts, for example, starting from the forecast values ​​of major macroeconomic indicators of one country, that is, forecasting the trends of the entire economy, the future performance of individual companies and the trends of stock prices are quantitatively measured. Macroeconomic index calculation means that calculates the future macroeconomic index of a country based on the major macroeconomic exogenous indicators assumed by the device user, recognizing the issue of providing a suitable stock price forecasting device as a forecasting tool. And, the individual company management index calculation means for calculating the future management index of the individual company based on the calculated future macroeconomic index, and the calculated future management index and the future macroeconomic index based on the calculated future macroeconomic index. A stock price forecasting device (see Patent Document 1), which is characterized by being provided with an individual company stock price calculation means for calculating the future stock price of the individual company, has been proposed.
Prior art literature
Patent documents
[0004]
Patent Document 1: Japanese Unexamined Patent Publication No. 10-3465
Outline of the invention
Problems to be solved by the invention
[0005]
　When forecasting the future, there are issues due to personality, such as the accuracy differs depending on the skills and experience of each person in charge at the financial institution. On the other hand, from the point of view of the person in charge, the load associated with the work is not small, and it is easy to create a need to improve work efficiency and accuracy. In addition, financial institutions tend to be concerned about the trends of other companies regarding such future forecasts, but there is no mechanism to obtain the corresponding information and effectively utilize it in their own business.
[0006]
　Therefore, an object of the present invention is to provide a technique for improving the accuracy and efficiency of future prediction associated with a predetermined business of a financial institution.
Means to solve problems
[0007]
　The prediction business support device of the present invention that solves the above problems is a storage device that stores information on a predetermined index value and information on a predetermined event used in a financial institution, and stores the index value and each information of the event from the storage device. The index value is read out, and the correlation analysis between the index value and the information of the event is executed by inputting each of the information, an event having a predetermined influence on the index value is specified, and the index value is calculated based on the information of the event. A process of generating an estimated regression equation, a process of generating a question screen including an interface capable of selectively answering the trend prediction of the event constituting the variable of the regression equation, and a process of generating the question screen are predetermined. It is distributed to the terminal of a financial institution, the answer of the trend prediction is obtained from the terminal via the question screen, the value of the answer is input to the variable of the corresponding event in the regression equation, and the prediction of the index value is predicted. It is characterized by including a calculation device that executes a process of calculating a value, a process of outputting information on the predicted value to a predetermined device of a predetermined financial institution, and a process of executing the process.
[0008]
　Further, in the prediction business support method of the present invention, an information processing device provided with a storage device that stores information on a predetermined index value and information on a predetermined event used by a financial institution can use the index value and each information of the event. By reading from the storage device and inputting each of the information, a correlation analysis of the index value and the information of the event is executed, an event having a predetermined effect on the index value is identified, and the event is based on the information of the event. A process of generating a regression equation for estimating an index value, a process of generating a question screen including an interface capable of selectively answering a trend prediction for the event constituting the variable of the regression equation, and the question. The screen is distributed to the terminal of the predetermined financial institution, the answer of the trend prediction is obtained from the terminal via the question screen, the value of the answer is input to the variable of the corresponding event in the regression equation, and the index is described. It is characterized in that a process of calculating a predicted value of a value and a process of outputting information on the predicted value to a predetermined device of a predetermined financial institution are executed.
Effect of the invention
[0009]
　According to the present invention, it is possible to improve the accuracy and efficiency of future prediction associated with a predetermined business of a financial institution.
A brief description of the drawing
[0010]
[Fig. 1] Fig. 1 is a network configuration diagram including a prediction work support device in the present embodiment.
[Fig. 2] Fig. 2 is a diagram showing a hardware configuration example of the prediction work support device of the present embodiment.
[Fig. 3A] It is a figure which shows the composition example of the market condition information (interest rate) in this embodiment.
[Fig. 3B] It is a figure which shows the composition example of the market condition information (stock price index) in this embodiment.
[Fig. 3C] It is a figure which shows the composition example of the market condition information (exchange rate) in this embodiment.
[Fig. 4] Fig. 4 is a diagram showing a configuration example of event information in the present embodiment.
FIG. 5 is a diagram showing a configuration example of regression equation information in the present embodiment.
[Fig. 6A] Fig. 6A is a diagram showing a configuration example of question management information in this embodiment.
[Fig. 6B] Fig. 6B is a diagram showing a configuration example of response management information in this embodiment.
FIG. 7 is a diagram showing a configuration example of an analyst report in this embodiment.
FIG. 8 is a flow chart showing a procedure example 1 of the prediction work support method of the present embodiment.
FIG. 9 is a diagram showing a configuration example of range information in the present embodiment.
FIG. 10 is a diagram showing a screen example 1 in the present embodiment.
[Fig. 11A] It is a figure which shows the composition example of the forecast market condition information (interest rate) in this embodiment.
[Fig. 11B] It is a figure which shows the composition example of the forecast market condition information (stock price index) in this embodiment.
[Fig. 11C] It is a figure which shows the composition example of the forecast market condition information (exchange rate) in this embodiment.
FIG. 12 is a diagram showing a screen example 2 in the present embodiment.
FIG. 13 is a diagram showing a screen example 3 in the present embodiment.
FIG. 14 is a diagram showing a screen example 4 in the present embodiment.
FIG. 15 is a flow chart showing a procedure example 2 of the prediction work support method of the present embodiment.
FIG. 16 is a diagram showing a conceptual example of a histogram in the present embodiment.
FIG. 17 is a flow chart showing a procedure example 3 of the prediction work support method of the present embodiment.
Mode for carrying out the invention
[0011]
--- Network Configuration ---
　Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an example of a network configuration including the prediction business support device 100 of the present embodiment. The prediction business support device 100 shown in FIG. 1 is an information processing device for improving the accuracy and efficiency of future prediction associated with a predetermined business of a financial institution.
[0012]
　Such a prediction business support device 100 is connected to the network 10 and is capable of data communication with the user terminal 200 and the information distribution server 300.
[0013]
　Of these, the user terminal 200 is a terminal operated by a person in charge of a financial institution. This person in charge is a person in charge of the ALM business, who wants to improve the efficiency and accuracy of his / her own business by using the service provided by the prediction business support device 100 of the present embodiment.
[0014]
　The information distribution server 300 is a server that distributes economic information such as official commissions, stock indexes, and exchange rates, and news information on politics such as political changes and economic policies. It can be assumed that the information distribution server 300 is operated by, for example, a company that provides an economic information distribution service.
[0015]
--- Hardware configuration example --- The hardware configuration of
　the prediction business support device 100 is as follows. In the prediction business support device 100, a storage device 101 composed of an appropriate non-volatile storage element such as a hard disk drive, a memory 103 composed of a volatile storage element such as RAM, and a program 102 held in the storage device 101 are stored in the memory 103. It is provided with at least a CPU 104 (computing device) that performs various determinations, calculations, and control processes while performing integrated control of the system itself by reading and executing, and a communication device 105 that is connected to the network 10 and is responsible for communication processing with other devices. ..
[0016]
　Further, the storage device 101 described above stores a program 102 for implementing a function associated with the prediction work support method of the present embodiment. The program 102 also includes a correlation analysis engine 110. The correlation analysis engine 110 is a program that performs correlation analysis, and an existing one may be appropriately adopted. However, instead of the correlation analysis engine 110, for example, an artificial intelligence that provides the same function may be adopted, and the configuration for performing the correlation analysis is not limited.
[0017]
　Further, in the storage device 101, in addition to the above-mentioned program 102, the index value information 125, the event information 126, the regression equation information 127, the question management information 128, the answer management information 129, the analyst report 130, and the prediction index value Information 131, is stored. Details of this information will be described later.
[0018]
--- Data Structure Example ---
　Next, the databases used by the prediction business support device 100 of the present embodiment will be described. 3A to 3C show an example of market condition information (interest rate) 125A, market condition information (stock price index) 125B, and market condition information (exchange rate) 125C, which are index value information 125 in this embodiment. It should be noted that each value constituting the index value information 125 can be assumed to be received and stored by the prediction business support device 100 from the information distribution server 300.
[0019]
　Of these, the market information (interest rate) 125A shown in FIG. 3A is a table that stores the actual value of the interest rate, which is a kind of economic indicator, in chronological order.
[0020]
　The data structure is a collection of records consisting of data such as base date, interest rate sum name, type, currency, period, and rate, using the market information code as a key.
[0021]
　Of these, the market information code is an ID that uniquely identifies the type of interest rate. The base date is a value indicating when the interest rate is applied. The Japanese name and type of interest rate are values ​​indicating the Japanese name and type of the interest rate. The currency is a value indicating the applicable currency of the interest rate. The period is a value indicating the period until the maturity of the interest rate. The rate is a value indicating the value of the interest rate.
[0022]
　The market information (stock price index) 125B shown in FIG. 3B is a table in which the actual values ​​of the stock price index, which is a kind of economic index, are stored in chronological order.
[0023]
　The data structure is a collection of records consisting of data such as a base date, a stock index Japanese name, and a rate, using a market information code as a key.
[0024]
　Of these, the market information code is an ID that uniquely identifies the type of the stock price index. The base date is a value indicating the calculation time of the stock price index. The Japanese name of the stock price index is a value indicating the Japanese name of the stock price index. The rate is a value indicating the value of the stock price index.
[0025]
　The market information (exchange rate) 125C shown in FIG. 3C is a table in which the actual value of the exchange rate, which is a kind of economic indicator, is stored in time series.
[0026]
　The data structure is a collection of records consisting of data such as a base date, an exchange rate Japanese name, and a rate, using a market information code as a key.
[0027]
　Of these, the market information code is an ID that uniquely identifies the type of exchange rate. The base date is a value that indicates when the exchange rate is applied. The exchange rate Japanese name is a value indicating the Japanese name of the exchange rate. The rate is a value indicating the value of the exchange rate.
[0028]
　Subsequently, FIG. 4 shows an example of the event information 126 in this embodiment. The event information 126 is a table that stores information on events related to the economy in chronological order. It should be noted that each value constituting the event information 126 can be assumed to be received and stored by the prediction business support device 100 from the information distribution server 300.
[0029]
　The data structure is a collection of records consisting of data such as the date of occurrence of the event, the event sum name, and the value, using the event code that uniquely identifies the event as a key.
[0030]
　Of these, the date of occurrence is a value indicating the time of occurrence of the event. The event Japanese name is a value indicating the Japanese name of the event. The value is a value indicating the value of the event. This value is set to "0" if there is no economic turmoil, "1" if there is economic turmoil, and so on if the corresponding event, that is, the event is "presence or absence of economic turmoil". If the corresponding event, that is, the event, is (announcement of) the "economic growth rate", the value of the economic growth rate announced by the government agency is set.
[0031]
　Subsequently, FIG. 5 shows an example of the regression equation information 127 according to the present embodiment. The regression equation information 127 is a table storing the regression equation obtained by analyzing the above-mentioned index value information 125 and event information 126 with the correlation analysis engine 110 as inputs.
[0032]
　The data structure uses the market information code corresponding to the market information (index value) that is the calculation result of the regression equation as a key, and the application start date of the regression equation, the regression equation, and the variables and answer keys in the regression equation. It is a collection of records consisting of data such as the correspondence of.
[0033]
　Of these, the application start date is a value indicating the target time of the index value, which is the market information estimated by the regression equation. Further, in the regression equation, the output y is the value of the market condition information indicated by the market condition information code (example: the value of the interest rate), and is composed of a combination of the variable X for calculating the market condition information and its coefficient. .. For example, an equation such as y = 1.01, X1 + 0.002, X2 + 0.124, X3 + ... Can be assumed.
[0034]
　The variables X1, X2, ... X7 are variables corresponding to a predetermined event in the event information 126, and are "question contents" shown in the question management information 128 of FIG. 6A for each type of answer to the question related to the event. It is provided for each "option" of ". Therefore, in the above regression equation, for example, the variable "X1" is the "economic growth rate (increase)", the variable "X2" is the "economic growth rate (slight increase)", and the variable "X3" is the "economic growth rate". (Flat) ”, variable“ X4 ”is“ economic growth rate (slight decrease) ”, variable“ X5 ”is“ economic growth rate (decrease) ”, variable“ X6 ”is“ presence or absence of economic turmoil (yes) ” , The variable "X7" is a variable such as "presence or absence of economic turmoil (none)".
[0035]
　In terms of correspondence with the "answer key", the variable "X1" is "00101", the variable "X2" is "00102", the variable "X3" is "00103", and the variable "X4" is "00103". "00104", the variable "X5" corresponds to "00105", the variable "X6" corresponds to "00201", and the variable "X7" corresponds to "00202". This correspondence is specified in the "correspondence column between variables and answer keys" in the regression equation information 127.
[0036]
　Therefore, the content that the person in charge of the financial institution has answered to the question described later, that is, the answer key stored in the "answer number" column of the answer management information 129 corresponds to any of the above variables. is there.
[0037]
　In addition, the values ​​assigned to these variables correspond to the presence or absence of answers regarding the "choices" of the "question contents" corresponding to the variables. There was an answer, that is, the answer key is included in the "answer number" column of the answer management information 129, and the variable corresponding to the question content option is "1", and the question content option for which there was no answer The corresponding variable is "0".
[0038]
　For example, in a certain record of the answer management information 129, it is assumed that the answer keys "00101" and "00201" are included in the "answer number" column. In that case, the prediction business support device 100 substitutes "1" as the value of the variable "X1" corresponding to the answer key "000101", and also substitutes "1" for the variable "X6" corresponding to the answer key "000201". "1" is substituted as the value, and "0" is substituted as the value of the other variables "X2" to "X5" and "X7" (because there is only one answer to each question).
[0039]
　In addition, the coefficient to be multiplied by each variable in the above-mentioned regression equation for economic growth rate is determined by inputting the actual change in economic growth rate so far and the range assigned to each answer key (that is, each variable). It is a coefficient value calculated by the artificial intelligence of. Of course, the method for determining the coefficient is not limited to this, and it is sufficient if the existing technology can be applied to appropriately determine the coefficient. It should be noted that the above-mentioned artificial intelligence can be assumed to be prepared for use by the prediction business support device 100 of the present embodiment or to use the artificial intelligence analysis service provided in the network 10.
[0040]
　Subsequently, FIG. 6A shows an example of the question management information 128 in the present embodiment. The question management information 128 is a table that stores information on questions that ask a person in charge of a financial institution or the like about an event corresponding to a variable included in each regression equation in the regression equation information 127 described above.
[0041]
　The data structure is a collection of records composed of data such as the question content, answer key, and options of the question, using the question key that uniquely identifies the question as a key.
[0042]
　Of these, the question content is a value indicating the content to be asked to the person in charge in the question. The answer key is a key that uniquely identifies the content of the answer given by the person in charge to the question for each type. The options also indicate the branches of the answer that the person in charge can select, and are uniquely linked to the answer key described above.
[0043]
　For example, regarding the question content of the question key "001", "Please select your view on Japan's economic growth rate", the answers are "increase" in the answer key "00101" and "slight increase" in the answer key "00102". , The answer key "00103" is "flat", the answer key "00104" is "slightly reduced", and the answer key "00105" is "decreased".
[0044]
　Subsequently, FIG. 6B shows an example of the response management information 129 in this embodiment. The answer management information 129 is a table that stores the contents that the person in charge of the financial institution has answered to the question managed by the question management information 128 described above.
[0045]
　The data structure uses the bank code that uniquely identifies the financial institution that sent the response as a key, and the registration date of the response by the person in charge of the financial institution, the ID of the selected scenario, the name of the scenario, and the response number. , And a collection of records consisting of data such as the scenario selection process.
[0046]
　Of these, the scenario ID is an ID for uniquely identifying the answer information input by the user with the bank code + serial number or the like. Further, the answer number is information composed of the answer keys of the options specified by the above-mentioned question management information 128 and selected by the person in charge. In addition, the scenario selection process stores information that describes the reason for the information that the person in charge used as the basis for selecting the scenario.
[0047]
　Subsequently, FIG. 7 shows an example of the analyst report 130 according to the present embodiment. The analyst report 130 is a table that stores an analyst report created by a predetermined financial analyst. Each analyst report stored in the analyst report 130 is provided in advance by a predetermined financial institution.
[0048]
　The data structure is a collection of records consisting of data such as the title of the analyst report and the contents of the report, with the registration date of the analyst report as a key. Of these, the report title is the title of the analyst report, and the report content is the content of the analyst report.
[0049]
--- Flow example 1 ---
　The actual procedure of the prediction work support method in this embodiment will be described below with reference to the figure. Various operations corresponding to the predictive business support method described below are realized by a program read by the predictive business support device 100 into the memory 103 and executed. Then, these programs are composed of codes for performing various operations described below.
[0050]
　FIG. 8 is a flow chart showing a processing procedure example 1 of the prediction work support method in the present embodiment. Here, prior to the flow, it is assumed that, for example, a notification requesting a predicted value for a predetermined index value is received from a user terminal 200 of a certain financial institution.
[0051]
　Therefore, the prediction business support device 100 acquires the corresponding information from the index value information 125 of the storage device 101 with respect to the index value (hereinafter, the required index value) requested by the above notification (s100). When the required index value is "interest rate", the information extracted from the index value information 125 is market information (interest rate) 125A.
[0052]
　Subsequently, the prediction business support device 100 accesses the storage device 101 and acquires the event information 126 (s101). In addition to the event information 126, the index value information 125 other than the above-mentioned required index value, that is, the market condition information (stock price index) 125B and the market condition information (exchange rate) 125C may be acquired. This is because other index values ​​can be assumed as events that affect the required index value, in addition to the event indicated by the event information 126.
[0053]
　Next, the prediction business support device 100 executes the correlation analysis between the required index value and the event or other index value by giving each information obtained in the above-mentioned s100 and s101 to the correlation analysis engine 110 as an input. s102). This correlation analysis itself is similar to the existing one.
[0054]
　Further, the prediction business support device 100 identifies an event that has a predetermined effect on the required index value, that is, an event or another index value by the above-mentioned correlation analysis, and estimates the required index value based on the information of the event or the like. Generate an expression (s103). This regression equation assumes a linear equation in which the value of event information 126 and other index values ​​are variables X1 to Xn, where the required index value is "y". As the method for generating the regression equation here, an existing one may be appropriately adopted.
[0055]
　The prediction business support device 100 has, among the required index values ​​"interest rate" and the event information 126, which are input to the algorithm for generating the regression equation such as artificial intelligence when generating the regression equation described above. For example, each value of the event information 126 is classified according to the variables X1 to X7 already described, that is, each concept of the option for the question. Therefore, the prediction business support device 100 holds the range information 1000 (FIG. 9) in the storage device 101.
[0056]
　In the example of the range information 1000 shown in FIG. 9, the economic growth rate "is a value of the economic growth rate corresponding to each event such as" increase "," slight increase ", ... Regarding the variables X1 to X5, that is, the" economic growth rate ". In the case of "increase", the value is "6% or more", in the case of "slight increase", the value is "less than 6% 3% or more", and in the case of "flat", the value is "less than 3% 0% or more". In the case of "slight decrease", the value is defined as "-3% or more and less than -6%", and in the case of "decrease", the value is defined as "-6% or more".
[0057]
　Similarly, in this range information 1000, the values ​​corresponding to the variables X6 to X7, that is, the events such as "yes" and "no" related to "presence or absence of economic turmoil" are set to "1" in the case of "yes". , In the case of "none", the value is defined as "0".
[0058]
　Therefore, when the prediction business support device 100 performs the above-mentioned regression equation generation, for example, each value of the "economic growth rate" which is the event information 126 to be processed is collated with the above-mentioned range information 1000, and the value is changed. , "Increase", "Slight increase", "Flat", "Decrease".
[0059]
　For example, when the economic growth rate "0.038", that is, "3.8%" of the date of occurrence "1980101" is collated with each provision of "economic growth rate" in the range information 1000, it is "less than 6% and 3% or more". Because it is in the range of ", it can be judged that the economic growth rate is" slight increase ".
[0060]
　The forecasting business support device 100 is the event information 126 that has undergone such a determination, and each value of the economic growth rate (here, either "increase", "slight increase", ...) And the value of the presence or absence of economic turmoil (here). Gives artificial intelligence the rate of the corresponding index value for each year, that is, the "interest rate" (eg, the Japanese name of the interest rate "JPY Tibor1M" of the market information code "10001"). The combination of variables and coefficients in the regression equation will be specified and the regression equation will be generated.
[0061]
　Further, the prediction business support device 100 stores the regression equation generated in s103 in the regression equation information 127 (s104).
[0062]
　Subsequently, the prediction business support device 100 includes an interface capable of selectively answering the trend prediction of the events and other index values ​​constituting the variables with respect to the regression equations generated and stored in the above-mentioned s103 and s104. A question screen is generated (s105).
[0063]
　For example, suppose that the regression equation is related to the required index value "interest rate", and the variables are "economic growth rate" and "presence or absence of economic turmoil". In this case, the forecasting business support device 100 stores a pull-down menu for answering the trend forecast of "economic growth rate" from the options and a pull-down menu for answering the trend forecast of "presence or absence of economic turmoil" from the options as questions. A predetermined screen format held in 101 is set in advance, and a question screen is generated.
[0064]
　FIG. 10 shows a specific example of the question screen 900 in this embodiment. This question screen 900 is referred to as a "scenario registration screen" for convenience. Here, the configuration includes an input interface for asking the person in charge of the financial institution for an answer for each item of the scenario name 901, category 902, question 903, and answer history 904.
[0065]
　Of these, the scenario name 901 is a field in which the person in charge freely inputs a name indicating the economic situation at the time when the required index value is estimated. On the other hand, the category 902 is a free entry field in which the person in charge can freely enter an economic situation such as an optimistic scenario or a pessimistic scenario. The optimistic scenario assumes the economic situation without large-scale economic turmoil, and the pessimistic scenario assumes the economic situation with large-scale economic turmoil.
[0066]
　In addition, question 903 is a pull-down that includes options such as increase, slight increase, flat, slight decrease, and decrease regarding the trend prediction of the variable "economic growth rate" among the trends of events and other index values ​​that compose the variables of the regression equation. It is composed of a menu 9031 and a pull-down menu 9032 including options such as yes or no regarding the trend prediction of the variable "presence or absence of economic turmoil".
[0067]
　Further, the answer process 904 is an input field for freely inputting the process of the answer selected by the person in charge in such an input interface.
[0068]
　Subsequently, the prediction business support device 100 delivers the above-mentioned question screen 900 to the user terminal 200 of the financial institution with which a predetermined agreement has been concluded in advance (s106).
[0069]
　On the other hand, the user terminal 200 of the financial institution displays the question screen 900, and in each interface of the scenario name 901, the division 902, and the question 903 indicated by the question screen 900, the person in charge of ALM business, the financial analyst, etc. Will be obtained and returned to the forecasting business support device 100 as a response to the trend forecast.
[0070]
　For example, the name of the scenario 901 is "optimistic scenario (no assumption of large-scale economic turmoil)", the category 902 is "optimistic scenario", the answer to "QA1" in question 903 is "increase", and the answer to "QA2". "Yes", as the response process 904, "As for the outlook for the economic growth rate, based on the view of the Japanese economic outlook by the International Monetary Fund (IMF), it was judged that the real economic growth rate of Japan this year will increase slightly. It is assumed that a value such as "........." is returned from the user terminal 200 to the prediction business support device 100.
[0071]
　In this case, the prediction business support device 100 receives the above-mentioned answer value from the user terminal 200, inputs the value to the corresponding event or index value variable in the regression equation generated in s103, and predicts the index value. Is calculated (s107).
[0072]
　In this case, in the prediction business support device 100, the regression equation used in this case is, for example, interest rate y = 1.01 × X1 + 0.002 × X2 + 0.124 × X3 + 0.212 × X4 + 0.001 × X5 + 0.212 × X6 + 0.001. × X7, and the answer regarding "economic growth rate" is "increase", that is, the answer key "00101" is obtained (that is, the variable X1 is "1"), and the answer regarding "presence or absence of economic turmoil" Is "Yes", that is, if the answer key "00201" is obtained (that is, the variable X6 is "1"), it can be calculated that the interest rate y = 1.01 × 1 + 0.212 × 1 = 1.222.
[0073]
　The prediction business support device 100 that has calculated the predicted value of the required index value in this way is associated with the scenario and stores the information in the response management information 129. Further, it is assumed that the value indicated by the information is stored in the rate column in the prediction index value information 131. 11A to 11C show forecast market information (interest rate) 131A, forecast market information (stock index) 131B, and forecast market information (exchange rate) 131C, which are examples of forecast index value information 131.
[0074]
　As shown here, the forecast index value information 131 is a record composed of the same items as the index value information 125 such as the base date, Japanese name, and rate, using the market information code corresponding to the required index value as a key. It is a collection of one answer).
[0075]
　Subsequently, the prediction business support device 100 outputs the information of the predicted value calculated in s107 to the requested (or predetermined financial institution) user terminal 200 (s108), and ends the process.
[0076]
　At the time of this output, the prediction business support device 100 may receive a condition designation from the user terminal 200 and output information according to the condition designation.
[0077]
　For example, as shown in the scenario search screen 1500 of FIG. 12, the prediction business support device 100 receives the designation of each condition of the bank name, the scenario classification, and the registration period for the scenario desired by the person in charge, and is based on the specified conditions. The search is executed with the answer management information 129. Then, a record including keywords such as "optimism" and "pessimism" indicated by the scenario classification in the scenario name can be specified. The record specified here is displayed on the scenario search screen 1500 as a "scenario list".
[0078]
　The person in charge of operating the user terminal 200 browses the scenario search screen 1500 and selects a scenario for which detailed information is to be confirmed from the above-mentioned "scenario list". Then, the prediction business support device 100 obtains information about the corresponding scenario from the response management information 129, generates a scenario reference screen 1600 including the information, and returns the information to the user terminal 200.
[0079]
　As illustrated in FIG. 13, the scenario reference screen 1600 includes scenario basic information, scenario selection process, and question / answer information. Further, when a predetermined operation is received on the scenario reference screen 1600, the prediction business support device 100 shifts the screen to the scenario reference screen 1700 illustrated in FIG. 14, where the market information classification, the market information, and the graph thereof. Is displayed. Both are obtained from the response management information 129.
[0080]
　With such a configuration, it is possible to search and confirm information such as trend judgment of economic indicators at other financial institutions, which has been difficult to know until now. Such search / confirmation operations can be performed by each financial institution with respect to other financial institutions, that is, the above-mentioned information and the like can be shared between financial institutions. From the point of view of the person in charge of ALM business mentioned above, a function that can improve the efficiency and accuracy of business, such as ensuring the validity of the judgment of one's own bank or making minor corrections as appropriate by referring to the judgment tendency of other banks. It becomes. As a result, it becomes possible to further improve the accuracy and efficiency of future forecasts associated with the prescribed operations of financial institutions.
[0081]
--- Flow Example 2 ---
　Subsequently, the processing related to the above-mentioned range information 1000 will be described. FIG. 15 is a flow chart showing a processing procedure example 2 of the prediction work support method in the present embodiment. Here, an example of generating the above-mentioned range information 1000 will be described.
[0082]
　When the prediction business support device 100 generates the above-mentioned question screen 900, the information about the event specified in s102 as an event affecting the required index value, that is, the value or index value information of the target event in the event information 126 The rate of the target index value at 125 is extracted from the event information 126 and the index value information 125 (s200). Here, as an example, it is assumed that the event "economic growth rate" is specified as an event that affects the required index value, and each value of this "economic growth rate" is extracted from the event information 126.
[0083]
　Subsequently, the forecasting business support device 100 sets each value of the "economic growth rate" extracted in s200 according to the number of options in the pull-down menu 9031 of the "economic growth rate" on the question screen 900 for each size. Classify and generate a histogram 1300 showing the distribution of the values ​​of the event "economic growth rate" (s201).
[0084]
　In the case of the histogram 1300 shown in FIG. 16, the value of the economic growth rate is "-6% or less", "-6% or more and less than -3%", corresponding to the number of the above-mentioned choices being "5". The value of "economic growth rate" is classified into each class of "0% or more and less than 3%", "3% or more and less than 6%", and "6% or more", and the frequency of appearance is taken on the vertical axis. It has become.
[0085]
　If the economic growth rate is "-6% or less", the frequency of appearance is "5", and if the frequency of appearance is "-6% or more and less than -3%", the frequency of appearance is "12", "0% or more and 3%". In the case of "less than", the frequency of appearance is "21", in the case of "3% or more and less than 6%", the frequency of appearance is "31", and in the case of "6% or more", the frequency of appearance is "20". ing.
[0086]
　Next, the forecasting business support device 100 links the range information of the value of the event "economic growth rate" corresponding to each class of the histogram 1300 generated as described above to the corresponding options in the pull-down menu 9031. It is stored in the range information 1000 (s202).
[0087]
　In this case, the forecasting business support device 100 has "increase", "slight increase", "flat", "slight decrease", "decrease", in order of the magnitude of the value of "economic growth rate" in each of the above-mentioned classes. Link the range of values ​​in the class. For example, "6% or more" is associated with the option "increase", "3% or more and less than 6%" is associated with the option "slight increase", and "flat" is associated with the option "flat". "0% or more and less than 3%" is associated, "-6% or more and less than -3%" is associated with the option "slight decrease", and "-6% or less" is associated with the option "decrease". To tie.
[0088]
　In generating the question screen 900, the prediction business support device 100 may compare the frequencies of each of the classes of the histogram 1300 and determine the classes whose frequencies are equal to or higher than the predetermined reference. In other words, the class that appears too infrequently is excluded because it cannot be a valid option on the question screen 900.
[0089]
　In this case, the prediction business support device 100 specifies the number and contents of the pull-down menu options on the question screen 900 as the number of classes specified in this determination and the range of the value of the event in the class.
[0090]
--- Flow example 3 ---
　Next, the tuning process of the regression equation will be described. FIG. 17 is a flow chart showing a processing procedure example 3 of the prediction work support method in the present embodiment.
[0091]
　The prediction business support device 100 has generated a regression equation as described above, but it may be uncertain whether the accuracy is good from the beginning.
[0092]
　Therefore, the prediction business support device 100 sets the actual value of the corresponding event (corresponding to the variable of the regression equation) at a predetermined time as an input to the generated regression equation, and sets a model value of the index value. Calculate (s300). The above-mentioned real value corresponds to the value of an economic event such as "economic growth rate" in the event information 126. That is, although the regression equation was generated through correlation analysis or the like, the actual value of the event information 126 (the value converted to "1" or "0" according to the choice of the corresponding event in the judgment based on the range information 1000). Is calculated as a model value when the index value is assigned to a variable.
[0093]
　For example, suppose that the target regression equation is interest rate y = 1.01 × X1 + 0.002 × X2 + 0.124 × X3 + 0.212 × X4 + 0.001 × X5. Also assume that the actual economic growth rate for a certain year was "4%". At this time, the forecasting business support device 100 collates this actual economic growth rate "4%" with the range information 1000, which corresponds to "slight increase" among the existing options and corresponds to the variable "X2". Identify what you are doing. Further, the prediction business support device 100 substitutes “1” for the variable “X2” of the regression equation described above, and calculates the interest rate y = 0.002 × 1 = 0.002, that is, 2% as the model value.
[0094]
　Subsequently, the prediction business support device 100 compares the above-mentioned model value with the predicted value calculated in s107, and determines whether the deviation exceeds a predetermined reference (s301).
[0095]
　For example, suppose that the predicted value is "1.8%", the model value is "2%", and the reference value of dissociation is "1.5%". In that case, it can be determined that the deviation between the predicted value and the model value is “0.2%” and the reference value of the deviation does not exceed “1.5%”.
[0096]
　As a result of the above determination, if the deviation does not exceed the predetermined standard (s302: n), the prediction business support device 100 ends the process.
[0097]
　On the other hand, when the deviation exceeds a predetermined standard as a result of the above determination (s302: y), the prediction business support device 100 inputs information on a new event that is not used for generating the regression equation from the input device 105 or storage. Obtained from device 101 (s303).
[0098]
　For example, in the correlation analysis of the regression equation, when the information given as input to the correlation analysis engine 110 is the value of "economic growth rate", the value of "presence or absence of economic turmoil" is stored in the event information 126 of the storage device 101. Get from.
[0099]
　Subsequently, the prediction business support device 100 adds the information of the new event obtained in s303 to the event information originally adopted (in the case of the above example, the “economic growth rate”) of the correlation analysis engine 110. The regression equation is regenerated as an input (s304). The content of the correlation analysis itself is the same as that of s102.
[0100]
　Further, the prediction business support device 100 stores the regression equation regenerated in s304 in the regression equation information 127 (s305), and ends the process.
[0101]
　The prediction business support device 100 is measuring to continuously tune the regression equation and improve its accuracy by repeatedly executing each of the processes of s300 to s305 at predetermined periods.
[0102]
　Similar to Flow Example 1, the regression equation tuned in this way is used in processing such as generating a question screen and calculating a predicted value based on the answer value of the trend prediction, and presenting a highly accurate result to the person in charge. Leads to.
[0103]
　In this embodiment, the ALM business is assumed as the case to be processed by the prediction work support device 100, but the type of the case is not limited to this.
[0104]
　For example, it can be assumed that the forecasting business support method of the present embodiment is applied to cases such as real estate loans, corporate loans, and deposit balances.
[0105]
　For example, in the case of a real estate loan, the loan contract rate and the loan balance can be assumed as index values, and events such as long-term interest rate, short-term interest rate, land route price, and average income can be assumed as events that affect them. In the case of a corporate loan, the loan contract rate and the loan balance can be assumed as index values, and events such as long-term interest rate, short-term interest rate, corporate tax rate, and corporate sales can be assumed as events that affect the loan contract rate and loan balance. In the case of the deposit balance, the deposit balance and the fixed deposit acquisition rate can be assumed as index values, and events such as long-term interest rate, short-term interest rate, personal income, and corporate sales can be assumed as events affecting this.
[0106]
　According to the prediction business support device of the present embodiment, it is possible to improve the accuracy and efficiency of future prediction associated with the predetermined business of the financial institution.
[0107]
　The description herein reveals at least the following: That is, when generating the question screen, the arithmetic unit classifies the information of the event according to the magnitude of the value of the specified event according to the number of choices in the selectable interface, and generates a histogram. Then, in the process of further executing the process of associating the range information of the value of the predetermined event corresponding to each class in the histogram with the option and holding it in the storage device and generating the regression equation, the specified influence. By collating the information of the event that causes the event with the range information and determining which of the options the information corresponds to, the information of the event is converted into the value of any of the options. It may be acquired, the acquired value and the index value are given to a predetermined machine learning algorithm, and a regression equation is generated by specifying a combination of variables and their coefficients in the regression equation.
[0108]
　According to this, it is possible to convert values ​​such as the economic growth rate used when generating the regression equation into options on the question screen, that is, values ​​to be assigned to the corresponding variables in the regression equation, and give them to the machine learning algorithm. , It is possible to efficiently generate a simple regression equation that directly corresponds to the choices on the question screen. As a result, it becomes possible to further improve the accuracy and efficiency of future forecasts associated with the prescribed operations of financial institutions.
[0109]
　In the prediction business support device of the present embodiment, the calculation device converts the actual value of the event at a predetermined time by collating it with the range information as an input to the variable for the regression equation. One of the values ​​is set, the model value of the index value is calculated, the model value is compared with the predicted value, and if the deviation exceeds a predetermined standard, it is not used in the generation of the regression equation. Regarding the process of acquiring information on a new predetermined event from an input device or storage device, adding the information on the new event, and then regenerating the regression equation, and the events constituting the variables of the regenerated regression equation. , A process of generating a question screen including an interface capable of answering the trend forecast in a selectable manner, and delivering the question screen to a terminal of a predetermined financial institution, and answering the trend forecast from the terminal via the question screen. Is acquired, the value of the answer is input to the variable of the corresponding event in the regression equation, the process of calculating the predicted value of the index value, and the information of the predicted value is output to the predetermined device of the predetermined financial institution. It may be the one that executes the process.
[0110]
　According to this, it is possible to regenerate the regression equation in the direction of eliminating the discrepancy between the predicted value and the model value, and gradually improve the accuracy of the regression equation, that is, the accuracy of the predicted value. As a result, it becomes possible to further improve the accuracy and efficiency of future forecasts associated with the prescribed operations of financial institutions.
[0111]
　In the prediction business support device of the present embodiment, when generating the question screen, the arithmetic unit compares the frequencies of each of the classes, determines the class whose frequency is equal to or higher than a predetermined reference, and determines the class with the option. Is specified as the number of classes specified in the determination and the range of values ​​of the predetermined event in the class, and the question screen including the interface corresponding to the number and contents of the specified options is generated. It may be a thing.
[0112]
　According to this, among the values ​​of events such as the economic growth rate and the official discount rate, for example, the range of values ​​that rarely occur is excluded, and the range of values ​​whose frequency of occurrence is appropriate is divided and the class is set. , It will be possible to adopt it as a proof of choice. As a result, it becomes possible to further improve the accuracy and efficiency of future forecasts associated with the prescribed operations of financial institutions.
[0113]
　In the prediction business support device of the present embodiment, the computing device receives a request for presenting information about another financial institution different from the financial institution from the terminal of the predetermined financial institution, and has already obtained the information regarding the other financial institution. The process of outputting at least one of the information of the response value of the trend prediction and the predicted value of the calculated index value to the terminal of the other financial institution may be further executed. ..
[0114]
　According to this, it becomes possible for each financial institution to search and confirm information such as trend judgment of economic indicators and the like in other financial institutions, which has been difficult to know until now. That is, the above-mentioned information and the like can be shared between financial institutions. From the perspective of the person in charge of ALM operations, it is possible to improve the efficiency and accuracy of operations by referring to the judgment trends of other banks and ensuring the validity of their own judgments or making minor corrections as appropriate. As a result, it becomes possible to further improve the accuracy and efficiency of future forecasts associated with the prescribed operations of financial institutions.
[0115]
　In the prediction business support method of the present embodiment, when the information processing apparatus generates the question screen, the event is determined for each magnitude of the value of the specified event according to the number of choices in the selectable interface. The information in the above is classified to generate a histogram, and the process of associating the range information of the value of the predetermined event corresponding to each class in the histogram with the option and holding it in the storage device is further executed, and the regression equation is obtained. In the process of generating, the information of the identified affected event is collated with the range information, and it is determined which of the options the information corresponds to, so that each of the information of the event is selected from the options. Assuming that the value converted to one of the values ​​is acquired, the acquired value and the index value are given to a predetermined machine learning algorithm, the combination of the variables and their coefficients in the regression equation is specified, and the regression equation is generated. May be good.
[0116]
　In the prediction business support method of the present embodiment, the above-mentioned option in which the information processing apparatus collates the actual value of the event at a predetermined time with the range information and converts the regression equation as an input to a variable. One of the values ​​is set, the model value of the index value is calculated, the model value is compared with the predicted value, and if the deviation exceeds a predetermined standard, it is used to generate the regression equation. A process of acquiring information on a new predetermined event from an input device or a storage device, adding information on the new event, and then regenerating a regression equation, and an event constituting a variable of the regenerated regression equation. With respect to, a process of generating a question screen including an interface capable of selectively answering the trend forecast, and delivering the question screen to a terminal of a predetermined financial institution, and transmitting the trend forecast from the terminal via the question screen. A process of acquiring an answer, inputting the value of the answer into the variable of the corresponding event in the regression equation, calculating the predicted value of the index value, and outputting the information of the predicted value to a predetermined device of a predetermined financial institution. You may execute the process to be performed.
[0117]
　In the prediction business support method of the present embodiment, when the information processing device generates the question screen, the frequency is compared for each of the classes, and the class whose frequency is equal to or higher than a predetermined standard is determined. The number and contents of choices are specified as the number of classes specified in the determination and the range of values ​​of the predetermined event in the classes, and the question screen including the interface according to the number and contents of the specified choices is generated. You may do.
[0118]
　In the prediction business support method of the present embodiment, the information processing apparatus receives a request for presenting information about another financial institution different from the financial institution from the terminal of the predetermined financial institution, and has already obtained the information regarding the other financial institution. , The process of outputting at least one of the respective information of the response value of the trend prediction and the predicted value of the calculated index value to the terminal of the other financial institution may be further executed.
Description of the sign
[0119]
10 Network
100 Forecasting business support device
101 Storage device
102 Program
103 Memory
104 CPU (Calculator)
105 Communication device
110 Correlation analysis engine
125 Index value information
125A Market information (interest rate)
125B Market information (stock index)
125C Market information (exchange rate) )
126 Event information
127 Regressive information
128 Question management information
129 Answer management information
130 Analyst report
131 Predictive index value information
131A Predicted market information (interest rate)
131B Predicted market information (stock index)
131C Predicted market information (exchange rate)
200 Users Terminal
300 Information distribution server
1000 Range information
The scope of the claims
[Claim 1]
　A storage device that stores information on a predetermined index value used by a financial institution and information on a predetermined event, and the
　index value and each information of the event are read out from the storage device, and the index value and the previous index value are input by inputting the information. A process of performing a correlation analysis with the information of the above-mentioned event, identifying an event having a predetermined influence on the index value, and generating a regression equation for estimating the index value based on the information of the event, and a process of the regression equation. A process of generating a question screen including an interface capable of selectively answering the trend prediction of the event constituting the variable, and a process of delivering the question screen to the terminal of a predetermined financial institution and passing through the question screen. A process of acquiring the answer of the trend prediction from the terminal, inputting the value of the answer into the variable of the corresponding event in the regression equation, calculating the predicted value of the index value, and predetermining the information of the predicted value.
　A forecasting business support device including a process for outputting to a predetermined device of a financial institution and a computing device for executing the process .
[Claim 2]
　When
　generating the question screen , the arithmetic unit classifies the information of the event according to the magnitude of the value of the specified event according to the number of choices in the selectable interface, and generates a histogram. The
　specified influence is exerted in the process of generating the regression equation by further executing the process of associating the range information of the value of the predetermined event corresponding to each class in the histogram with the option and holding it in the storage device. By collating the event information with the range information and determining which of the options the information corresponds to, the information of the event converted into the value of any of the options is acquired. , giving the the acquired value as the index value to a predetermined machine learning algorithm, in which by specifying a combination of variables and their coefficients in the regression equation to generate the regression equation,
　it in claim 1, wherein The described forecasting business support device.
[Claim 3]
　The calculation device sets,
　as an input to a variable, a value of any of the above options obtained by collating the actual value of the event at a predetermined time with the range information and converting the regression equation. A model value of the index value is calculated, the model value is compared with the predicted value, and if the deviation exceeds a predetermined standard, information on a new predetermined event that is not used for generating the regression equation is input from the input device. Regression equations are regenerated after acquiring information from the storage device and adding information on the new
　event, and trend predictions can be selected for the events that make up the variables of the regenerated regression equation. The process of generating a question screen including a regressive interface, the question screen is distributed to the terminal of a predetermined financial institution, the answer of the trend prediction is obtained from the terminal via the question screen, and the value of the answer is calculated. The process of inputting into the variable of the corresponding event in the regression equation and calculating the predicted value of the index value and the process of outputting the information of the predicted value to a predetermined device of a predetermined financial institution
　are executed. 2. The forecasting business support device according to claim 2.
[Claim 4]
　When
　generating the question screen , the arithmetic unit compares the frequencies of each of the classes, determines the class whose frequency is equal to or higher than a predetermined standard, and specifies the number and contents of the options by the determination. The
　claim is characterized in that it identifies the number of classes and the range of values ​​of the predetermined event in the class, and generates the question screen including the interface according to the number and contents of the specified options. Item 2. The predictive business support device according to item 2.
[Claim 5]
　The computing device
　receives a request for presenting information about another financial institution different from the financial institution from the terminal of the predetermined financial institution, and has already obtained the value of the answer to the trend forecast and the value of the answer to the other financial institution.
　The forecasting business according to claim 1 , wherein at least one of the calculated predicted values ​​of the index value is further executed to be output to the terminal of the other financial institution. Support device.
[Claim 6]
　An information processing device provided with a storage device that stores information on a predetermined index value and information on a predetermined event used by a financial institution
　reads out the index value and each information of the event from the storage device and inputs the respective information. A process of executing a correlation analysis between the index value and the information of the event, identifying an event having a predetermined effect on the index value, and generating a regression equation for estimating the index value based on the information of the event. When,
　wherein with respect to the events that constitute a regression equation of variables, the process of generating a question screen including an interface which can be answered selectable trends predicted,
　to deliver the question screen to the terminal of a predetermined financial institution, The process of acquiring the answer of the trend prediction from the terminal via the question screen, inputting the value of the answer into the variable of the corresponding event in the regression equation, and calculating the predicted value of the index value, and the
　above.
　A forecasting business support method characterized by executing a process of outputting forecasted value information to a predetermined device of a predetermined financial institution .
[Claim 7]
　When the information processing device
　generates the question screen , the information processing device classifies the information of the event according to the magnitude of the value of the specified event according to the number of choices in the selectable interface, and generates a histogram. In the process of further executing the process of associating the range information of the value of the predetermined event corresponding to each class in the histogram with the option and holding it in the storage device
　and generating the regression equation, the specified influence is exerted. By collating the information of the event to be exerted with the range information and determining which of the options the information corresponds to, the information of the event is converted into the value of any of the options. The
　sixth aspect of claim 6 is characterized in that the acquired value and the index value are given to a predetermined machine learning algorithm, and a combination of variables and their coefficients in the regression equation is specified to generate a regression equation . Predictive business support method.
[Claim 8]
　The information processing apparatus
　sets, as an input to the variable, the value of any of the above options converted by collating the actual value of the event at a predetermined time with the range information with respect to the regression equation. A model value of the index value is calculated, the model value is compared with the predicted value, and if the deviation exceeds a predetermined standard, information on a new predetermined event that is not used for generating the regression equation is input. Or, for the process of regenerating the regression equation after acquiring the information of the new event from the storage device and
　the event that constitutes the variable of the regenerated regression equation, the trend prediction can be selected and answered. The process of generating a question screen including a possible interface, the question screen is distributed to the terminal of a predetermined financial institution, the answer of the trend prediction is obtained from the terminal via the question screen, and the value of the answer is obtained. the regression type variable corresponding event in formula, the processing to calculate the predicted value of the index value, and outputting the information of the predicted value in a predetermined unit of predetermined financial institution,
　characterized in that to perform the The forecasting business support method according to claim 7.
[Claim 9]
　When the information processing device
　generates the question screen, the frequency is compared for each of the classes, the class whose frequency is equal to or higher than a predetermined standard is determined, and the number and contents of the options are determined by the determination.
　7. Claim 7 is characterized in that the number of specified classes and the range of values ​​of the predetermined event in the class are specified, and the question screen including the interface according to the number and contents of the specified options is generated. Predictive business support method described in.
[Claim 10]
　The information processing apparatus
　receives a request for presenting information about another financial institution different from the financial institution from the terminal of the predetermined financial institution, and has already obtained the value of the response of the trend forecast and the response to the other financial institution.
　The forecasting business support method according to claim 6, further executing a process of outputting at least one of the calculated information of the index value predicted value to the terminal of the other financial institution. ..