JP5680144B2 - Prediction device, prediction method, and computer program - Google Patents

Description

  The present invention relates to a prediction apparatus, a prediction method, and a computer program for predicting a value related to an event that changes in time series.

  Economic forecasts such as product demand forecasts and sales forecasts are extremely important in examining the direction and strategy in company management. Then, how to connect the predicted demand to the plans of each department such as sales, inventory, production, distribution, and development is a management issue. Furthermore, not only forecasts for economic events such as demand forecasts and sales forecasts, but forecasting future events using information that has changed over time is an important issue in various fields. .

  Various time-series analysis methods have been proposed as methods for predicting subsequent changes in time-series events such as the number of products demanded or the number of products sold. Examples of such an analysis method include multivariate analysis such as multiple regression analysis and T method (for example, Patent Document 1 and Non-Patent Document 1). Furthermore, various application proposals have been proposed for these analysis methods (Non-Patent Document 2, etc.).

Japanese Patent No. 3141164

Edited by Kazuo Tatebayashi, Shoichi Teshima, Ryoko Hasegawa, “Introductory MT System”, Nikka Giren Shuppansha, December 2008 Yukiya Masuda, “Research of T Method Considering Nonlinear Components”, Proceedings of 17th Quality Engineering Research Conference, p. 422-425, 2009

  The T method is not originally a method for prediction of time series data. There is no particular concept of the time axis, and it is a method for selecting which factor has the most influence on the value of an event related to a large number of factors. When the T method is used as a time series data prediction method, the relationship between the item value of each item and the signal value in a certain period is specified, and the signal value in the next period is predicted. For this reason, there has been a problem that estimation errors in the item values of the respective items overlap to deteriorate the prediction accuracy of the signal values.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a prediction device, a prediction method, and a computer program capable of improving prediction accuracy by reflecting a tendency of a signal to change over time. .

The prediction device according to the present application records a value related to an event that changes in time series and an item value of each item that includes each of a plurality of related events related to the event in a time series as a record. A value related to the event in a first period within a period recorded in the recording means, and an item value related to the related event in a second period before the first period. based on the strength of the correlation, a prediction device for predicting the value of the event in the prediction period, among the plurality of items item value in the recording means are recorded, is recorded before type recording means extracting means for item value in it is time to extract an item which does not include a period in which a zero, a plurality of the total value obtained by summing the item value of each item for each item of the segment as a new item value Items and individual items extracted by the extraction means are Setting means for setting as an item to be used for prediction of values related to a value according to the event, using the item value of the item of which setting unit has set, deriving means for deriving the strength of the correlation, and the A prediction unit is provided that calculates a predicted value related to the event in the prediction target period based on the strength of the correlation derived by the deriving unit.

  In the prediction device of the present application, the derivation unit is configured to calculate a factor effect value indicating a correlation strength of each item with respect to a change in the value related to the event, and to calculate the event based on the calculated factor effect value. Item selection means for selecting an item to be used for prediction of such a value.

  In the prediction device of the present application, the item selection unit calculates, for each item, the strength of correlation with the prediction target for a plurality of items selected in order from the largest factor effect value, and the calculated number of items. Means for determining the number of items based on the strength of each correlation, and means for selecting items for the determined number of items in descending order of the factor effect value.

The prediction method of the present application is a recording in which a value related to an event that changes in time series and an item value of each item that includes each of a plurality of related events related to the event are associated and recorded in time series A value relating to the event in the first period within the period recorded in the recording means by the computer provided with the means, and an item relating to the related event in the second period before the first period. A prediction method for predicting a value related to the event in a prediction target period based on the strength of correlation with a value, wherein the computer includes a plurality of items whose item values are recorded in the recording unit , before Symbol field values within a period recorded in the recording means extracts an item that does not contain a period during which a zero sum new items obtained by summing the item value of each item for each item of the segment Multiple items as values and individual extracted Eyes, and set as an item to be used for prediction of values related to the event, using the item value of the items set to the value according to the event, the derived correlation strengths, the derived correlation A predicted value related to the event in the prediction target period is calculated based on strength.

The computer program of the present application records a time series in which values related to events that change in time series are associated with item values of each item having each of a plurality of related events related to the events as items. And a value related to the event in the first period within the period recorded in the recording means, and an item related to the related event in the second period before the first period. A computer program for predicting a value related to the event in a prediction target period based on the strength of correlation with a value, wherein the computer includes a plurality of items whose item values are recorded in the recording means , before Symbol extracting means item value within a period recorded in the recording means to extract an item which does not contain a period during which a zero, a new total value obtained by summing the item value of each item for each item of the segment Items A plurality of items and, the individual items extracted by the extracting unit, setting means for setting as an item to be used for prediction of values related to the event, and a value relating to the event, the setting unit has set Deriving means for deriving the strength of the correlation using the item value of the item, and prediction for calculating a predicted value related to the event in the prediction target period based on the strength of the correlation derived by the deriving means It functions as a means.

  In the present application, from the recording means for recording the value related to the event that changes in time series and the item value of each item having each of the related events related to the event as items, in time series, An item corresponding to each of a plurality of categories corresponding to the attribute of the related event and an item not including a period in which the item value is zero are extracted, and prediction is performed using the item values of these items.

  According to this application, for example, when performing sales forecasts for products that are deployed globally, items are classified according to attributes such as regions, major customers, and small customers, and sales results for individual customers are added as items. By doing so, prediction accuracy can be improved.

It is explanatory drawing which shows the signal used by T method, and the example of the content of each item. It is explanatory drawing which shows an example of the proportionality constant (beta) and SN ratio (eta) (square ratio) of each item used by T method. It is the figure which showed the actual signal value of each member or the actual value of the standardized signal value, and the comprehensive estimated value calculated | required about each member in list format. It is a factor effect figure which shows the S / N ratio of the comprehensive estimated value of each item in T method. It is a graph which shows the influence with respect to the factor effect of each item in T method. It is explanatory drawing which illustrates a time difference model notionally. It is a graph which shows notionally the processing for determining the number of item selection. It is a graph which shows a response | compatibility with the number of item selections, and SN ratio of a comprehensive estimated value. It is a block diagram which shows the structure of the prediction apparatus which concerns on this Embodiment. It is a flowchart which shows an example of the prediction process procedure by a prediction apparatus. It is a flowchart which shows the process sequence of variable conversion. It is a flowchart which shows the calculation procedure of SN ratio of a comprehensive estimated value. It is explanatory drawing which shows an example of an item. It is a graph which shows the transition of the sales volume to a certain individual customer. It is a graph which shows the relationship between the signal value after variable conversion, and an item value. It is a graph which shows transition of the sales volume to another individual customer. It is a graph which shows the relationship between the signal value after variable conversion, and an item value. It is a graph which shows transition of the sales volume to another individual customer. It is a graph which shows the relationship between the signal value after variable conversion, and an item value. It is a graph which shows the prediction result at the time of employ | adopting the sales volume to an individual customer as an item. It is explanatory drawing which shows an example of the item by a sales package model. It is a graph which shows the prediction result at the time of employ | adopting the sales volume to a sales package as an item. It is explanatory drawing which shows an example of the item employ | adopted by this application. It is a graph which shows the prediction result at the time of employ | adopting a sales package model and the sales volume to an individual customer as an item. It is a graph which shows transition of the prediction accuracy accompanying the difference in item contents. It is a graph which shows the prediction result of a single year. It is a graph which shows the prediction result of a single year.

The present invention will be specifically described with reference to the drawings showing the embodiments thereof.
The prediction method of the present application is based on a method using an MT system (for example, T method) that has been used in the past, in order to predict a change over time of a value (signal) related to an event that changes in time series, A time difference model is introduced. In the present application, item values each having a plurality of related events that may cause a signal are recorded in time series in association with the signal, and prediction is made based on the strength of correlation between the signal and the item value. One of the features is to select signals and items to be used.

  In addition, the following prediction methods are implemented by calculation means, such as a computer (prediction apparatus 1 mentioned later) which can read each data from the recording means which has recorded the data used for prediction.

First, an outline of the T method which is the basis of the prediction method of the present application will be described. FIG. 1 is an explanatory diagram showing examples of signals used in the T method and the contents of each item. In FIG. 1, a member is an index indicating data for each unit period, and is indicated as 1, 2,..., N. The value relating to the event is recorded as a signal value M. “Item 1”, “Item 2”,..., “Item k” indicate each of a plurality of related events related to the signal value M. X ₁₁ , X ₁₂ ,..., X _1k indicate item values of each item related to the signal value M ₁ of member 1. For example, if the member has monthly electricity usage, the monthly electricity usage (W) is the signal value M, and the item values are “month”, “temperature”, “wind speed”, “precipitation ( Record the variables that would be related to the ups and downs of electricity usage, such as “monthly average”, “sunshine hours (monthly average)”, “maximum temperature (monthly average)”, “minimum temperature (monthly average)”.

  In addition, although the actual value may be used for the signal value M and each item value X, it is preferable to use them after normalization using a calculation means. In the normalization method, for example, the calculation means subtracts the average value of the item value of each item (for example, the average value of the monthly average temperature of all members to be used) from the item value X of each item (for example, the monthly average temperature). There is a way to keep it. Thereby, when the item value is plotted against the signal value (referred to as a unit space), the prediction formula can be expressed as a straight line passing through the origin.

  In the T method, a specific formula that specifies the relationship between the signal value M of each member and the related multi-variable item value is obtained, the item value of the next unit period to be predicted is predicted, and the predicted item A value is applied to a specific formula to predict. However, the item data is multivariate. For example, there may be 12 members and there are 20 items (k = 20). In such a case, it is difficult to obtain a specific expression for predicting the next signal value. In the T method, for each of a large number of items, a value indicating the strength of the factor effect with respect to signal value fluctuation is evaluated, and a value indicating the factor effect is weighted. Thereby, by selecting effective items, it is possible to perform prediction with sufficiently high accuracy by using the selected items without using data of all items.

  Therefore, in the T method, the calculation means calculates the proportionality constant β and the SN ratio η (square ratio) for each item using the members having the signal value M by applying the following formulas 1 and 2. To do. The S / N ratio is a value indicated by using the reciprocal of dispersion as shown in Equation 2, is the sensitivity of the signal value for each item, and indicates the strength of correlation between each item and the signal value.

  In addition, the above-mentioned Formula 1 and Formula 2 are formulas for obtaining the proportionality constant β and the SN ratio η (square ratio) for Item 1. The calculation means performs the same calculation as for item 1 for item 2 to item k. FIG. 2 is an explanatory diagram showing an example of the proportionality constant β and SN ratio η (square ratio) of each item used in the T method. In FIG. 2, the proportionality constant β and SN ratio η (square ratio) for each item calculated by applying Equation 1 and Equation 2 described above to each item are shown in a table format.

  Next, using the proportionality constant β and the SN ratio η (square ratio) for each item, an estimated value of the output for each item is obtained for each member by the computing means. For the i-th member, the estimated output value of item 1 can be expressed by the following equation 3. Similarly, the estimated value of the output from item 2 to item k is obtained by the calculation means.

  Thereby, the calculating means can derive the comprehensive estimation formula (Formula 4) as a prediction formula indicating the relationship between the item value and the total estimated value of the signal value. However, the comprehensive estimation formula using all items (1 to k) with respect to the signal value to be predicted does not have the highest prediction accuracy. Therefore, in order to increase the contribution to the influence on the prediction target and increase the prediction accuracy, an appropriate combination of items is selected from all items by the calculation means.

Therefore, the calculation means calculates an overall estimated value using the SN ratios η ₁ , η ₂ ,..., Η _k (square ratio), which is the estimation accuracy for the estimated value of each item, as a weighting coefficient. Therefore, the total estimated value of the i-th member can be expressed by the following Equation 4.

  FIG. 3 is a diagram showing, in a list form, the actual signal values of each member or the actual values of the standardized signal values and the total estimated values obtained for each member. Then, using the actual value and the total estimated value of each member obtained as shown in FIG. 3, the computing means calculates the SN ratio η (db) of the total estimated value for each item by the following equation 5. .

  In the T method, a method of obtaining a value called a factor effect value by using the total estimated value obtained for each item as described above, and selecting an item based on the factor effect value is used. FIG. 4 is a factor effect diagram showing the SN ratio of the overall estimated value of each item in the T method, and FIG. 5 is a graph showing the influence of each item on the factor effect in the T method.

In FIG. 4, the horizontal axis indicates items to be selected, and the vertical axis indicates the SN ratio of each item by taking the SN ratio of the comprehensive estimated value. In FIG. 4, for each item, the SN ratio of the total estimated value for each data including the data of the item, that is, the strength of the correlation with the signal value is shown on the left side, and each data excluding the data of the item is shown on the right side. The SN ratio of the comprehensive estimated value with respect to is shown. In the example shown in FIG. 4, there are 36 items, and there are 2 ³⁶ combinations of selections. The computing means derives the SN ratio of the prediction target for one or a plurality of items for each of these combinations. And a calculating means calculates the average value of SN ratio of the combination containing the said item used as object, and the average value of S / N ratio of the combination which does not contain the said item. In FIG. 4, the average value of the SN ratio including the data of the item calculated in this way is shown on the left side, and the average value of the SN ratio not including the item is shown on the right side for each item.

  In FIG. 5, the horizontal axis indicates items to be selected, and the vertical axis indicates the factor effect value, and the degree of the factor effect value for each item is indicated. The factor effect value on the vertical axis in FIG. 5 is the left side of the SN ratio on the right side (excluding items) of the SN ratio of the overall estimated value of each item indicating the strength of the correlation of the prediction target in FIG. Of the SN ratio (including items), that is, a value obtained by subtracting the right SN ratio from the left SN ratio. That is, for each item, the degree of effect on the S / N ratio when the item is included is not shown. Therefore, an item for which the calculated factor effect value is positive indicates that the SN ratio of the comprehensive estimated value is increased by using the item. In a method called a two-sided T method, only items having such a positive factor effect value are selected, and analysis by the T method as described above is performed.

  When prediction is performed using the T method (two-sided T method) as described above, the actual prediction accuracy still does not increase. Therefore, the inventor first applied the concept of a time difference model to the above-described T method, and performed conversion processing taking into account the nonlinearity between the signal value and the item value. Instead of the method of selection, a method of maximizing the SN ratio of the comprehensive estimated value is used.

Next, the time difference model introduced by the prediction method of the present application will be described.
The T method is not originally a method for prediction of time series data. There is no particular concept of the time axis, and it is a method for selecting which factor has the most influence on the value of an event related to a large number of factors. When the T method is used as it is in the prediction method, as described above, the relationship between the item value and the signal value for each item of each member is specified, and the signal value of the next member is predicted. Therefore, in order to predict the signal value of the next member, the item value for each item of the next member must be predicted. It is easily assumed that the estimation error of the item value of each item at this time overlaps and the prediction accuracy of the signal value is deteriorated. In addition, when predicting the next signal value based on time-series data, especially in fields such as economic forecasting or sales forecasting, there should be signs of future events. is there. Accordingly, the inventor decided to associate the association between the signal value and the item value for each item as described above not by the members of the same period but by the ones shifted by a predetermined time. Specifically, the computing means that implements the prediction method of the present embodiment associates the signal value after a predetermined period with the item value for each item in a certain period, and calculates Equations 1 to 5 between these values. And the signal value after a predetermined period is estimated (predicted) based on the item value for each item in the present (most recent) period using the comprehensive estimation formula obtained as a result.

FIG. 6 is an explanatory diagram conceptually illustrating the time difference model. FIG. 6 shows the passage of time from the left to the right in the figure. Each rectangle in the lower part of FIG. 6 indicates an item value of each item for each year, and each rectangle in the upper part indicates a signal for each year at the same time as each item. The time difference model, rather than associating the item value of the signal and each item in the same period, with respect to the signal M _i, item values of the items of predetermined period before _{X i-t 1, X i} -t 2, ... , X _{i−t k} (t = 1 year in the example of FIG. 6).
In the example of the time difference model shown in FIG. 6, the signal of one year is configured to correspond to the item value of the previous year. However, these periods are not limited to the periods shown in FIG. It is possible.

In the present embodiment, for example, in order to predict a signal (unknown signal) in 2005, which is the prediction target period, signals (known signals) for each year from 2002 to 2004, and from 2001 to 2003, The time difference model that associates the items of each year (known items) is introduced. In the prediction method of the present application, the calculation means calculates the signal value M ₁ and the item values X ₁₁ , X _{12 ,.} The T method is applied by applying the above relationship. The signal value of the prediction target period is obtained by inputting the item values of the respective items before the predetermined period (item values from January to December 2004 in the example of FIG. 6) into the comprehensive estimation formula (Formula 4). That is, in the time difference model in the prediction method of the present embodiment, the computing means specifies the relationship between the item values of a plurality of items and the signal values after a predetermined period, thereby determining the future signal from the past or current item value. It is possible to predict the value.

  In the prediction method of the present embodiment, the best prediction scheme is adopted in addition to the time difference model described above. In the best prediction scheme, items are subjected to second-order variable transformation for logarithmically transformed signals, dispersion is stabilized for both signals and items, and predicted values are selected by item selection that maximizes the total estimated S / N ratio in the T method calculation. Is the way to get.

  In the following, variable conversion performed in the best prediction scheme will be described. In the conventional T method, for each item, a straight line that passes through the zero point is set for the relationship between the item data and the signal value, and weighting based on the deviation from this straight line (numericalization as an SN ratio) is performed. In such a linear transformation, items that have less variation with respect to a straight line are items that contribute to changes in the signal value, and items that have more variation, that is, items that have no correlation between the change in the item value and the change in the signal value, This item does not contribute to the change in value. However, the item values of all items do not necessarily have a straight line, that is, a linear relationship in relation to the signal value. Therefore, in the best prediction scheme, logarithmic transformation is performed on the signal, and the item is subjected to secondary variable transformation on the logarithmically transformed signal.

  Next, the item number selection process will be described. FIG. 7 is a graph conceptually showing a process for determining the number of item selections. First, the time difference model is applied, and the signal-to-variable conversion signal value and the item value for each item are applied to calculate the SN ratio of the overall estimated value of each item by applying Equations 1 to 5. Then, as described with reference to FIGS. 4 and 5, the calculation means calculates a factor effect value for each item from the SN ratio of the comprehensive estimated value. As shown in FIG. 7, the computing means first sets the minimum value of the factor effect value for each item as an initial threshold value. The computing means selects an item whose factor effect value is equal to or greater than a threshold value, and calculates the SN ratio of the comprehensive estimated value with respect to the signal value (for example, all 1 to n). If the threshold is the initial value, all items are selected. Next, the calculation means sets a value obtained by adding a predetermined value to the threshold value as the next threshold value, similarly selects an item whose factor effect value is equal to or greater than the threshold value, and calculates the SN ratio of the comprehensive estimated value with respect to the signal value. . The calculation means repeats such processing until the threshold reaches the maximum value of the factor effect value, that is, reaches the line shown as MAX in FIG. The ratio can be calculated for each number of items. The calculation means sets the initial value of the threshold value to the maximum value MAX or more, decreases the threshold value by a predetermined value, selects an item that is a factor effect value above each threshold value, and calculates the SN ratio of the comprehensive estimated value. You may make it go.

  FIG. 8 is a graph showing the correspondence between the number of item selections and the SN ratio of the overall estimated value. The graph shown in FIG. 8 shows the transition of the SN ratio of the total estimated value with respect to the number of item selections, with the SN ratio (db) of the total estimated value on the vertical axis and the number of item selections on the horizontal axis. . Among the S / N ratio values, the open square is the S / N ratio of the overall estimated value when an item is selected using the two-sided T method of selecting an item having a positive factorial effect value. Open diamonds indicate the SN ratio of the overall estimated value when all items are selected. White circles indicate the number of items with the maximum total SN ratio. As shown in FIG. 8, in the method of selecting an item having a positive factor effect value, the SN ratio of the comprehensive estimated value does not become maximum. In the present embodiment, by implementing the method shown in FIG. 7, the number of items for which the SN ratio of the overall estimated value is the highest is determined, and the items for the determined number of items in descending order of the factor effect value. The method of selecting is adopted.

  As described above, in the present application, the time difference model is applied to the T method, and after performing the conversion process considering the nonlinearity between the signal value and the item value of each item, the SN ratio of the comprehensive estimated value is calculated. By applying an item selection method that maximizes the S / N ratio, the prediction accuracy is improved.

  Next, the prediction device 1 according to the present embodiment will be described. FIG. 9 is a block diagram showing a configuration of the prediction apparatus 1 according to the present embodiment. The prediction device 1 is a computer such as a personal computer or a server computer, and includes a control unit 10, a recording unit 11, a temporary storage unit 12, an input unit 13, and an output unit 14.

  The control unit 10 includes, for example, a CPU (Central Processing Unit). The control part 10 controls the operation | movement of each part of the hardware mentioned above based on the prediction program 2 demonstrated below, and exhibits the function as the prediction apparatus 1 which concerns on this Embodiment. In addition, the control part 10 has a function as a calculating means which implements the prediction method mentioned above.

  The recording unit 11 uses a nonvolatile memory such as a ROM (Read Only Memory) or a hard disk drive. The recording unit 11 may be an external hard disk drive, an optical disk drive, or another recording device connected via a communication network. That is, the recording unit 11 is a general term for one or a plurality of information recording media accessible from the control unit 10.

  The recording unit 11 records a prediction program 2 including various procedures for realizing the prediction method of the present embodiment. A part of the recording area of the recording unit 11 is used as a database (DB) 110 that records signal values and data of a plurality of items corresponding to the values (item values of each item). The control unit 10 can read and write signal values and item values of each item with respect to the database 110. The database 110 records the signal value of each member and the item value of each item in time series, for example, in the format shown in FIG.

  The temporary storage unit 12 is, for example, a nonvolatile memory such as a DRAM (Dynamic Random Access Memory) or an SRAM (Static Random Access Memory). The temporary storage unit 12 temporarily stores information generated by the processing of the control unit 10.

  The input unit 13 receives a user operation input using a keyboard, a mouse, or the like.

  The output unit 14 uses a display unit such as a liquid crystal monitor or a printing unit such as a printer, and outputs information processing results by the control unit 10.

  In the prediction apparatus 1 configured as described above, the control unit 10 executes a process based on the prediction program 2 to predict a signal value related to a future event.

  FIG. 10 is a flowchart illustrating an example of a prediction processing procedure performed by the prediction device 1. The control unit 10 receives input of the signal value and the item value of each item related thereto from the input unit 13, and records the received signal value and the item value for each item in the database 110 of the recording unit 11 (step S11). ). The signal values and item values for each item to be recorded in the database 110 may be input from other devices via the communication network as well as those input from the input unit 13 or other information recording media. May be input.

  The control unit 10 generates a time difference model based on the signal value recorded in the database 110 of the recording unit 11 and the item value for each item (step S12). The time difference model generated in step S12 is a model in which the item value for each item is associated with the signal value after a predetermined period of the item value of the item, as described with reference to FIG. That is, in step S12, the control unit 10 associates the signal value stored in time series with the item value for each item while shifting by a predetermined period.

  The control unit 10 performs variable conversion of the signal value and the item value for each item based on the relationship between the corresponding signal value and the item value for each item in the generated time difference model (step S13). The variable conversion processing procedure will be described in detail with reference to the flowchart of FIG.

  Based on the signal value after variable conversion and the item value of each item associated with the time difference model, the control unit 10 uses the above-described Equations 1 and 2 to calculate the proportionality constant β and the SN ratio η (2) for each item. (Multiplier ratio) is calculated (step S14).

  The control unit 10 calculates the estimated value of the output by Expression 3 for each member in the selected signal period using the proportional constant β and the SN ratio η (square ratio) for each item using Expression 3. Step S15).

  The control unit 10 calculates an overall estimated value using the SN ratio (square ratio), which is the estimation accuracy of the estimated value, as a weighting coefficient by using Equation 4 (step S16).

  Next, based on Equation 5, the control unit 10 calculates the SN ratio (db) of the overall estimated value of each item based on the signal value and the overall estimated value (step S17).

  The control unit 10 derives a factor effect value for each item (step S18). As described above, the factor effect value is the SN ratio of the overall estimated value with respect to the item value of each item including the item with respect to the SN ratio of the overall estimated value with respect to the item value of each item excluding the item. Calculated by obtaining the difference. The S / N ratio of the comprehensive estimated value is the strength of correlation of the item value of the item with respect to the signal value, and is a value expressed as a logarithm of a value proportional to the reciprocal of the variance.

  Next, the control unit 10 calculates, for each item number, the SN ratio of the comprehensive estimated value with respect to the item values of the plurality of items selected in descending order of the factor effect value (step S19). Details of the processing in step S19 have been described above with reference to FIG.

  The control unit 10 determines the number of items that maximizes the SN ratio based on the SN ratio of the comprehensive estimated value for each number of items (step S20).

  The controller 10 selects as many items as the number of items determined in step S20 (step S21). Then, the control unit 10 selects the item value among the item values for each item before the predetermined period associated with the prediction target period in the time difference model, that is, the item value for each item in the corresponding period closest to the prediction target period. Is applied to Equation 4 to calculate a predicted value (step S22). Β and η in Equation 4 are those calculated in step S14 (corresponding to the time difference model and calculated by the variable-converted signal value and the item value of each item). In step S <b> 22, the predicted value is output from the output unit 14 or recorded in the recording unit 11. If the signal value is normalized, it can be obtained by performing inverse transformation.

FIG. 11 is a flowchart showing a variable conversion processing procedure. The control unit 10 performs logarithmic conversion on each signal value M _i (i = 1, 2,..., N) (step S131). That is, the control unit 10 calculates m _i = log M _i when the signal value after logarithmic conversion is m _i (i = 1, 2,..., N).

Next, the control unit 10 performs secondary variable conversion on the item values X _i−tj (j = 1, 2,..., K) of each item associated with the time difference model with respect to the signal value m _i after logarithmic conversion. (Step S132). Specifically, the control unit 10 performs a quadratic _{regression on} the signal value m _i after logarithmic conversion with respect to the item value X _i−tj of each item, and converts the conversion coefficient by the number of items (coefficient of the quadratic term, primary term Coefficient, constant). And the control part 10 calculates _| requires the signal value corresponding to the item value Xi _-tj of each item on the regression curve calculated _| required for every item, and memorize _| stores the signal value as converted item value xi _-tj temporarily. The process memorize | stored in the part 12 grade | etc., Is performed.

Next, the control unit 10 performs variable conversion on the item value of the unknown item (step S133). In step S131, the converted signal value m _i can be calculated for the known signal, and in step S132, the item value x _i−tj of each item after conversion can be calculated for the known item. However, although the above-described unknown item exists as an item value, there is no corresponding unknown signal (predicted signal), and thus the item value after variable conversion cannot be obtained. For this reason, in this embodiment, the conversion coefficient obtained for the known item is diverted, and the variable conversion is performed for the item value of the unknown item.
The unknown item after the variable conversion is required in the step of calculating the predicted value in the above-described step S22. Therefore, the unknown item may be converted into a variable after the item is selected in step S21.

  FIG. 12 is a flowchart showing a procedure for calculating the SN ratio of the comprehensive estimated value. The control unit 10 sets a value equal to or less than the minimum value of the factor effect value derived in step S16 as an initial value of the threshold value (S171). Step S171 is a state in which an initial value is set in FIG. As shown in FIG. 7, the processing amount can be reduced by setting the minimum factor effect value as a threshold value.

  Next, the control unit 10 selects an item for which a factor effect value equal to or greater than the set threshold is calculated (S172). When the minimum value of the factor effect value is set as the initial value of the threshold value, all items are selected in the first step S172.

  Next, the control unit 10 calculates the strength of correlation of the prediction target for the data of the item selected in step S172, that is, the SN ratio of the comprehensive estimated value (S173), and calculates the SN ratio of the calculated comprehensive estimated value. It is recorded in the temporary storage unit 12 in association with the number of items (S174).

  Next, the control unit 10 determines whether or not the calculation of the SN ratio of the comprehensive estimated value for the selected item is completed (S175). Step S175 is a process for determining whether to end the repetitive process, and is a process for determining whether or not the calculation process of the SN ratio of the total estimated value for each item number has been completed for each item number. For example, if the set threshold value is greater than or equal to the maximum factor effect value, and if a factor effect value that exceeds the threshold value does not exist due to resetting of the threshold value described later, a comprehensive estimate for each calculated number of items Completion conditions such as when the number of values matches the number of items to be selected can be set as appropriate.

  When it is determined in step S175 that calculation of the SN ratio of the total estimated value for the selected item has not been completed and calculation of the SN ratio of the total estimated value for a smaller number of items is required (S175: NO), the control unit 10 resets the set threshold value to a value larger by a predetermined value (S176), proceeds to step S172, and repeats the subsequent processing.

  If it is determined in step S175 that the calculation of the SN ratio of the comprehensive estimated value for the selected item has been completed (S175: YES), the control unit 10 ends the process according to this flowchart.

  Here, the mode in which the initial value of the threshold value is set to be equal to or less than the minimum value of the factor effect value and the threshold value is reset so as to increase by a predetermined value is shown, but the reverse processing may be performed. . In other words, the initial value of the threshold is set to be equal to or greater than the maximum value of the factor effect value, the item for which the factor effect value equal to or greater than the set threshold is calculated is selected, and the threshold value is reset so as to decrease by a predetermined value. May be.

  Next, as a specific application example of the prediction method by the prediction device 1, assuming sales prediction of a product that is being developed globally, “total sales” is adopted as a signal, and “ The case where the “sales number” is adopted will be explained. As will be described later, the present application introduces the concept of “sales package model” in which not only the number of units sold to individual customers (manufacturers) but also the number of units sold in a certain region is classified based on the sales scale.

  First, as a comparative example, a case where attention is paid only to the number of individual customers sold will be described. FIG. 13 is an explanatory diagram showing an example of items. The item represented as “symbol” in FIG. 13 is recorded as monthly data and represents the number of small-sized general-purpose diesel engines sold. Each item represented by the symbols “A1” to “A3” indicates the number of units sold to the individual customers “A1” to “A3” belonging to the region A, respectively. Similarly, each item represented by symbols “B1” to “B3” indicates the number of units sold to individual customers “B1” to “B3” belonging to region B, and is represented by symbols “C1” to “C4”. Each item shown represents the number of units sold to individual customers “C1” to “C4” belonging to the region C, respectively.

  The items shown in FIG. 13 do not include all individual customers belonging to each region. As will be described later, when the period in which the number of units sold is zero within the recording period, the SN ratio of the item with respect to the signal is remarkably low or cannot be calculated from the definition of the T method. Was excluded.

  FIG. 14 is a graph showing the transition of the number of units sold to an individual customer, and FIG. 15 is a graph showing the relationship between the signal value and the item value after variable conversion. The horizontal axis of FIG. 14 indicates the year and month in time series, and the vertical axis indicates the number of units sold to an individual customer (for example, individual customer “A1”) that can be an item. The number of units sold is monthly data, and in the example shown in FIG. 14, 141 values are plotted between April 2000 and December 2011. A solid curve in the graph indicates a signal trend (trend curve). From the graph of FIG. 14, it can be seen that the number of units sold to the individual customer “A1” does not include a period in which the value is zero within the recording period, although it shows a periodic fluctuation.

  The graph shown in FIG. 15 shows the relationship between the signal values after variable conversion (total sales volume) and item values (sales volume to individual customer “A1”). In this embodiment, the signal value is logarithmically converted, and the item value is converted to a secondary variable with respect to the logarithmically converted signal value. In this example, as shown in FIG. 15, the item can be linearized with respect to the signal, and the calculated S / N ratio was as good as 15.18 db. For this reason, it was found that the number of units sold to the individual customer “A1” can be adopted as an item.

  FIG. 16 is a graph showing the transition of the number of units sold to another individual customer, and FIG. 17 is a graph showing the relationship between the signal value and the item value after variable conversion. The horizontal axis of FIG. 16 indicates the year and month in time series, and the vertical axis indicates the number of units sold to another individual customer (for example, individual customer “A4”). The number of units sold is monthly data, and in the example shown in FIG. 16, 141 values are plotted between April 2000 and December 2011. The graph of FIG. 16 indicates that the individual customer “A4” is a new customer, and an order has been started from April 2007.

  The graph shown in FIG. 17 shows the relationship between the signal values after variable conversion (total sales volume) and item values (sales volume to individual customer “A4”). In this embodiment, the signal value is logarithmically converted, and the item value is converted to a secondary variable with respect to the logarithmically converted signal value. In this example, as shown in FIG. 17, it was possible to linearize the item with respect to the signal, but the calculated S / N ratio was −3.94 db, which was extremely low. As a result, it was found that the number sold to the individual customer “A4” cannot be adopted as an item.

  FIG. 18 is a graph showing the transition of the number of units sold to another individual customer, and FIG. 19 is a graph showing the relationship between the signal value and the item value after variable conversion. The horizontal axis in FIG. 18 indicates the year and month in time series, and the vertical axis indicates the number of units sold to another individual customer (for example, individual customer “A5”). The number of units sold is monthly data, and in the example shown in FIG. 18, 141 values are plotted between April 2000 and December 2011. The graph of FIG. 18 shows that the individual customer “A5” was not a new customer but was lost during the period from March 2005 to March 2007.

  The graph shown in FIG. 19 shows the relationship between the signal values after variable conversion (total sales volume) and item values (sales volume to individual customer “A5”). In this embodiment, the signal value is logarithmically converted, and the item value is converted to a secondary variable with respect to the logarithmically converted signal value. In this example, as shown in FIG. 19, it was possible to linearize the item with respect to the signal, but it was impossible to calculate the SN ratio from the definition of the T method. As a result, it was found that the number sold to the individual customer “A5” cannot be adopted as an item.

  The individual customers that can be adopted as items as described above are determined, and the predicted value of the total sales is calculated using the sales numbers to the determined individual customers (for example, “A1” to “C4” shown in FIG. 13). Calculation is performed according to the procedure shown in the flowcharts of FIGS.

  FIG. 20 is a graph showing a prediction result when the number of units sold to individual customers is adopted as an item. The horizontal axis shows the year and month in chronological order, and the vertical axis shows the total sales. Values indicated by black circles from April 2000 to December 2008 indicate actual values (actual values). Moreover, the broken line after January 2004 has shown the result of the predicted value. The prediction is performed every year, and the graph shows the prediction results for five years from 2004 onward.

  When only individual customers (10 companies) are adopted as items, as shown in FIG. 20, there are some periods in which the predicted value deviates significantly from the actual value, and the predicted value reproduces the actual value well. I can not say. The SN ratio of the prediction shown in FIG. 20 was −79.9 db, and the average absolute percent error (MAPE) representing the accuracy evaluation of the predicted value was 61.7%.

  As a next comparative example, a case will be described below in which a “sales package model” in which the number of units sold in a certain region is sorted by paying attention to the sales scale is introduced.

  FIG. 21 is an explanatory diagram showing an example of items based on the sales package model. The items represented as “symbols” in FIG. 21 are recorded as monthly data, and represent the number of small land general-purpose diesel engines sold. In the example shown in FIG. 21, the sales package model is introduced, and the sales volume by region is divided into the sales volume for large customers and the sales volume for small customers. Respond to lost orders. For region D and region E, the total number of units sold to large customers and small customers was determined without sorting the number of units sold to large customers and the number sold to small customers.

  For example, items represented by symbols “A” to “C” indicate the total number of units sold to major customers in regions A to C, respectively. In addition, items represented by symbols “D” and “E” indicate the total number of units sold to major customers and small customers in region D and region E, respectively. Items represented by symbols “a” to “c” indicate the total number of units sold to small customers in regions A to C, respectively. Each of the above items corresponds to the sales package model described above.

  The sales numbers for the eight sales packages indicated by the symbols “A” to “E” and “a” to “c” in FIG. 21 are adopted as items, and signals (all Calculate the predicted value of the number of units sold. FIG. 22 is a graph showing a prediction result when the number of units sold to the sales package is adopted as an item. The horizontal axis shows the year and month in chronological order, and the vertical axis shows the total sales. Values indicated by black circles from April 2000 to December 2008 indicate actual values (actual values). Moreover, the broken line after January 2004 has shown the result of the predicted value. The prediction is performed every year, and the graph shows the prediction results for five years from 2004 onward.

  When sales volume for 8 sales packages is adopted as an item, as shown in FIG. 22, there is a period in which the forecast value deviates significantly from the actual value. It can be seen that no significant decrease can be predicted. Note that the predicted S / N ratio shown in FIG. 22 was −77.5 db, and the average absolute percent error (MAPE) was 70.3%.

  Next, application examples of the present application will be described. In this application, in addition to the number of sales to the eight sales packages described above, the number of sales to individual customers (10 companies) is adopted as an item. FIG. 23 is an explanatory diagram showing an example of items employed in the present application. The item represented as “symbol” in FIG. 23 is recorded as monthly data and represents the number of small-sized general-purpose diesel engines sold. In this example, the sales package model is introduced, and the sales volume by region is divided into sales volume for major customers and sales volume for small customers. Etc. For region D and region E, the total number of units sold to large customers and small customers was determined without sorting the number of units sold to large customers and the number sold to small customers.

  For example, items represented by symbols “A” to “C” indicate the total number of units sold to major customers in regions A to C, respectively. In addition, items represented by symbols “D” and “E” indicate the total number of units sold to major customers and small customers in region D and region E, respectively. Items represented by symbols “a” to “c” indicate the total number of units sold to small customers in regions A to C, respectively. Each of the above items corresponds to the sales package model described above.

  In the present application, in addition to the number of sales to eight sales packages represented by symbols “A” to “E” and “a” to “c”, the number of sales to individual customers is introduced as an item. In FIG. 23, items represented by symbols “A1” to “A3” indicate the number of units sold to individual customers “A1” to “A3” belonging to the region A, respectively. Similarly, items represented by symbols “B1” to “B3” indicate the number of units sold to individual customers “B1” to “B3” belonging to the region B, and are represented by symbols “C1” to “C4”. The item indicates the number of units sold to the individual customers “C1” to “C4” belonging to the region C, respectively.

  In this application, the number of units sold to eight sales packages and the number of units sold to individual customers of 10 companies are adopted as items, and the predicted value of the signal (total number of units sold) is calculated according to the procedure shown in the flowcharts of FIGS. calculate. FIG. 24 is a graph showing a prediction result when the sales package model and the number of units sold to individual customers are adopted as items. The horizontal axis shows the year and month in chronological order, and the vertical axis shows the total sales. Values indicated by black circles from April 2000 to December 2008 indicate actual values (actual values). Moreover, the broken line after January 2004 has shown the result of the predicted value. The prediction is performed every year, and the graph shows the prediction results for five years from 2004 onward.

  When sales volume for 8 sales packages and sales volume for individual customers of 10 companies are adopted as items, as shown in Fig. 24, there are values that deviate from actual values when looking at individual forecast values However, the overall trend can be predicted with high accuracy, and the trend has been successfully reproduced for the sudden decrease in sales volume since January 2008.

  The predicted S / N ratio shown in FIG. 24 was -64.0 db, and the mean absolute percent error (MAPE) was 19.9%. That is, the forecast was made by adopting only the number of sales to individual customers of 10 companies as an item (see FIG. 20), and the forecast was made by adopting only the number of sales to eight sales packages as an item. Compared with the results (see FIG. 22), the prediction method of the present application that employs both as items shows that the SN ratio and MAPE are greatly improved and the prediction accuracy is increased.

  FIG. 25 is a graph showing the transition of prediction accuracy due to the difference in item contents. In the graph shown in FIG. 25, the horizontal axis represents the item content, the left vertical axis represents the predicted SN ratio (db), and the right vertical axis represents the average absolute percentage error (MAPE). The transition of predicted SN ratio (db) and MAPE is shown. The white circles in the graph indicate the number of units sold into eight sales packages (basic package), the number of sales packages for eight sales packages and six individual customers, the number of sales packages for eight sales packages and ten individual customers, and The SN ratio of the prediction when the number of sales to 10 individual customers is estimated as an item is shown, and the black circles indicate their MAPE. The individual customers of 6 companies obtain the SN ratio of the desired characteristics from the MAPE of the predicted values for the five years from January 2004 to December 2008, and selectively extract those that have a positive factor effect. The items obtained by doing

  As is apparent from the graph shown in FIG. 25, the prediction accuracy was greatly improved by adding the number of units sold to the individual customers of the six companies with positive S / N ratios as positive items to the basic package. In addition, when the number of units sold to 10 individual customers not including the period when the value (monthly unit sold) is zero is added as an item to the basic package, the predicted SN ratio is the maximum and the MAPE is the minimum The prediction accuracy was the highest.

  FIG.26 and FIG.27 is a graph which shows the prediction result of a single year. The horizontal axis represents the number of item selections, the vertical axis represents the overall estimated S / N ratio (db), FIG. 26 shows the prediction results for fiscal 2004, and FIG. 27 shows the prediction results for fiscal 2008. In the graphs of FIG. 26 and FIG. 27, the triangle mark indicates the SN ratio of the overall estimation with the basic package as an item, and the square mark adds the number of units sold to the individual customers of the six companies described above to the basic package as an item. The SN ratio of the overall estimate is shown, and the circles indicate the SN ratio of the overall estimate obtained by adding the number of units sold to the individual customers of the 10 companies described above to the basic package as an item.

  Both the prediction results for FY 2004 and FY 2008 showed that the SN ratio of the overall estimate can be significantly increased by adding the number of units sold to 10 individual customers as an item to the basic package. That is, in the present application, it has been shown that the influence of the interaction can be avoided by using the item selection method for maximizing the total S / N ratio, and appropriate feature items can be taken into the sales package model to improve the prediction accuracy.

  The embodiment disclosed this time is to be considered as illustrative in all points and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the meanings described above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims. In addition, the technical features described in each embodiment can be combined with each other.

DESCRIPTION OF SYMBOLS 1 Prediction apparatus 2 Prediction program 10 Control part 11 Recording part 12 Temporary storage part 13 Input part 14 Output part 110 Database

Claims (5)

A recording unit that records a value related to an event that changes in time series and an item value of each item that includes each of a plurality of related events related to the event in a time-series manner; Based on the strength of correlation between the value related to the event in the first period within the period recorded in the means and the item value related to the related event in the second period before the first period. A prediction device for predicting a value related to the event in the prediction target period,
Wherein among the plurality of items item value is recorded in the recording means, extracting means for item value within a period recorded in the prior type recording means for extracting an item that does not include a period in which the zero,
Prediction of a value related to the event, a plurality of items with the total value obtained by summing the item value of each item for each item category as a new item value, and individual items extracted by the extraction means Setting means to set as an item used for
The value according to the event, using the item value of the item of which setting unit has set, deriving means for deriving the strength of the correlation, and based on the strength of the correlation derived by conductor detecting means, said prediction A prediction device comprising: a prediction unit that calculates a predicted value related to the event in a target period. The derivation means includes
Means for calculating a factor effect value indicating the strength of correlation of each item with respect to a change in value related to the event;
The prediction apparatus according to claim 1, further comprising: an item selection unit that selects an item to be used for predicting a value related to the event based on the calculated factor effect value. The item selection means includes
Means for calculating, for each number of items, the strength of correlation with a prediction target for a plurality of items selected in descending order of the factor effect value;
Means for determining the number of items based on the strength of correlation for each calculated number of items;
The prediction apparatus according to claim 2, further comprising: means for selecting items for the determined number of items in descending order of the factor effect value. By a computer provided with a recording means for recording a value related to an event that changes in time series and an item value of each item that includes each of a plurality of related events related to the event in time series And a strong correlation between the value related to the event in the first period within the period recorded in the recording means and the item value related to the related event in the second period before the first period. And a prediction method for predicting a value related to the event in the prediction target period,
The computer
Wherein among the plurality of items item value is recorded in the recording means, extracts an item the item value does not contain a period during which a zero within a period recorded in the prior type recording means,
As items used for prediction of values related to the above events, a plurality of items with the total value obtained by summing the item values of each item for each item category as new item values and the extracted individual items Set,
Using the value relating to the event and the item value of the set item, the strength of the correlation is derived,
A prediction method, comprising: calculating a prediction value related to the event in the prediction target period based on the strength of the derived correlation. In a computer provided with a recording unit that records a value related to an event that changes in time series and an item value of each item that includes each of a plurality of related events related to the event in time series And a strong correlation between the value related to the event in the first period within the period recorded in the recording means and the item value related to the related event in the second period before the first period. And a computer program for predicting a value related to the event in the prediction target period,
The computer,
Wherein among the plurality of items item value is recorded in the recording means, extracting means for item value within a period recorded in the prior type recording means for extracting an item that does not include a period in which the zero,
Prediction of a value related to the event, a plurality of items with the total value obtained by summing the item value of each item for each item category as a new item value, and individual items extracted by the extraction means Setting means to set as an item used for
The value according to the event, using the item value of the item of which setting unit has set, deriving means for deriving the strength of the correlation, and based on the strength of the correlation derived by conductor detecting means, said prediction A computer program for functioning as a prediction means for calculating a predicted value related to the event in a target period.

Images