JP5889002B2 - Demand forecasting device, demand forecasting method, and demand forecasting program - Google Patents

Description

  The present invention relates to a demand prediction apparatus, a demand prediction method, and a demand prediction program.

  Technologies for predicting demand are known. For example, Patent Document 1 discloses a technique for calculating a demand prediction function using a function of a method having the smallest error among a cumulative secondary method, a cumulative tertiary method, and a cumulative quaternary method.

JP 2007-293614 A

  The demand prediction technique predicts demand based on the actual value. However, the recent demand trend may vary from past performance trends. For this reason, with the above-described demand prediction technology, there is a case where the demand cannot be accurately predicted. It is an object of the present invention to provide a demand prediction apparatus, a demand prediction method, and a demand prediction program that can accurately predict demand.

  As one aspect, the demand forecasting apparatus stores, as one aspect, a storage unit that stores demand information in which a scheduled delivery date of a product for which an inquiry has been made, a scheduled number to be paid out, and an expectation level of an order are associated with each other, and the planned delivery date A demand prediction unit that predicts demand for the product by integrating a value obtained by multiplying the expected number by a coefficient corresponding to the length of the period and the expectation.

  As one aspect, the demand forecasting method is a demand forecasting method executed by the demand forecasting device, and stores the scheduled delivery date of the product that has been inquired, the planned number to be paid out, and the expectation of the order received. Obtaining demand information, and predicting the demand for the product by integrating a value obtained by multiplying the number corresponding to the length of the period until the scheduled delivery date and the coefficient corresponding to the expectation by the planned number. .

  As one aspect, the demand prediction program acquires, in the demand prediction device, demand information that holds a scheduled delivery date of a product that has been inquired, a planned number to be paid out, and an expectation degree of an order in association with each other; A step of predicting the demand for the product by integrating a value obtained by multiplying the planned number by a coefficient corresponding to a length of the period until the scheduled delivery date and the expected degree.

  The demand prediction apparatus, the demand prediction method, and the demand prediction program according to the present invention have an effect that demand can be accurately predicted.

FIG. 1 is a block diagram illustrating a demand prediction apparatus according to the present embodiment. FIG. 2 is a diagram illustrating an example of demand information. FIG. 3 is a diagram illustrating an example of dive demand information. FIG. 4 is a diagram for explaining a history of occurrence of past dive demand. FIG. 5 is a diagram illustrating an example of coefficient information. FIG. 6 is a diagram for describing setting of coefficient information. FIG. 7 is a diagram for describing setting of coefficient information. FIG. 8 is a diagram illustrating an example of coefficient pattern information. FIG. 9 is a diagram illustrating an example of selection information. FIG. 10 is a diagram for explaining the optimum combination search for selection information. FIG. 11 is a diagram for explaining the optimum combination search for selection information. FIG. 12 is a flowchart illustrating a processing procedure of the demand prediction process.

  Hereinafter, embodiments of a demand prediction apparatus, a demand prediction method, and a demand prediction program according to the present invention will be described in detail with reference to the drawings. Note that the present invention is not limited to the embodiments. Further, the constituent elements in this embodiment include those that can be easily assumed by those skilled in the art, those that are substantially the same, and those in the so-called equivalent range.

  First, the configuration of the demand prediction apparatus according to the present embodiment will be described. FIG. 1 is a block diagram illustrating a demand prediction apparatus according to the present embodiment. As shown in FIG. 1, the demand prediction device 10 includes a display unit 11, an input unit 12, a communication unit 13, a medium reading unit 14, a control unit 15, and a storage unit 16.

  The display unit 11 includes a display device such as a liquid crystal panel or an organic EL (Organic Electro-Luminescence) panel, and displays various information such as characters, symbols, and figures based on a control signal transmitted from the control unit 15. To do. The input unit 12 includes an input device such as a keyboard, and outputs a signal corresponding to an operation performed by the user on the input device to the control unit 15. The communication unit 13 controls transmission / reception of information with other devices based on a predetermined communication protocol. The medium reading unit 14 reads a program and data from a portable non-transitory storage medium such as an optical disk, a magneto-optical disk, or a memory card.

  The control unit 15 includes a CPU (Central Processing Unit) 151 that is an arithmetic device and a memory 152 that is a storage device, and implements various functions by executing programs using these hardware resources. Specifically, the control unit 15 reads out a program stored in the storage unit 16 and expands it in the memory 152, and causes the CPU 151 to execute an instruction included in the program expanded in the memory 152. The control unit 15 reads / writes data from / to the memory 152 and the storage unit 16 and controls the operation of the communication unit 13 and the like according to the execution result of the instruction by the CPU 151.

  The storage unit 16 includes a nonvolatile storage device such as a magnetic storage device or a semiconductor storage device, and stores various programs and data. The programs stored in the storage unit 16 include a demand prediction program 161. The data stored in the storage unit 16 includes demand information 162, dive demand information 163, coefficient information 164, coefficient pattern information 165, and selection information 166.

  1 may be stored in a storage medium that can be read by the medium reading unit 14. The program and the data that are stored in the storage unit 16 in FIG. Alternatively, all or part of the program and data that the storage unit 16 stores in FIG. 1 may be acquired from another device through communication by the communication unit 13.

  The demand prediction program 161 provides a function for creating a plan for predicting demand. The demand prediction program 161 includes a demand prediction unit 161a and an update unit 161b.

  The demand prediction unit 161a calculates the demand D in the designated period using the following formula (1).

  Here, Du is a demand that is not assumed in advance, and De is a demand that is assumed in advance. That is, the demand prediction unit 161a separately calculates a demand that is not assumed in advance and a demand that is assumed in advance, and calculates the demand in a specified period by summing them.

  The demand that is assumed in advance is a demand that is specifically predicted based on inquiries from sales activities, inquiries, and the like when the demand is predicted. The demand that is not assumed in advance is a demand that has no inquiry at the time when the demand is predicted. In the following description, a demand that is not assumed in advance may be referred to as a “jump demand”.

  Thus, by separately predicting demand that is assumed in advance and demand that is not assumed in advance, each demand can be predicted using a suitable method, and the accuracy of prediction is improved. A method for predicting each demand will be described later.

  The update unit 161b updates the coefficient used for the prediction by the demand prediction unit 161a. The updating unit 161b updates the coefficient based on the past inquiry and the actual result.

  The demand information 162 holds information about inquiries and results. FIG. 2 is a diagram illustrating an example of the demand information 162. The demand information 162 shown in FIG. 2 includes items such as ID, product number, type number, scheduled delivery date, planned number, expectation level, dive flag, payout date, payout number, and registration date. The ID item stores an identification number for identifying data. The product number field stores an identification number for identifying the product. In the type item, an identification number for identifying the type of product is stored.

  The scheduled delivery date field stores a value indicating the time when the product is scheduled to be dispensed. The number of products scheduled to be dispensed is stored in the item of scheduled number. In the expectation item, a value indicating a high possibility of order reception is stored. The value of the expectation item is determined based on the judgment of the person in charge who negotiates with the customer. In the present embodiment, the item of expectation stores one of “5”, “4”, “3”, “2”, and “1” in descending order of possibility of receiving an order.

  In the entry flag field, a value indicating whether or not an inquiry has been made in advance prior to the delivery of the product is stored. In this embodiment, when there is an inquiry in advance, “N” is stored in the item of the jump flag. If there is no inquiry in advance, “Y” is stored in the entry flag item. The demand information 162 may be configured to be able to determine whether or not there has been an inquiry in advance even without the entry flag item.

  The date of product delivery is stored in the item of delivery date. In the item of the number of payouts, the number of products paid out is stored. Until the product is paid out, the items of the payout date and the payout number are blank. If there is an inquiry but the product is not paid out because the contract is not established, the items of the payout date and the number of payouts remain blank. The date of data registration is stored in the item of registration date. In this embodiment, the date when the data is registered is treated as the date when the inquiry occurred.

  The jump-in demand information 163 holds information for predicting the jump-in demand. FIG. 3 is a diagram illustrating an example of the dive demand information 163. In the jump-in demand information 163, the predicted value of the jump-in demand generated in one month is stored for each part type and for each number of months until the payout date. For example, the data in the first row of the dive demand information 163 shown in FIG. 3 indicates that the dive demand generated in one month of the product of the type identified by “AAA” is paid out in the current month (after 0 month). Indicates that 18 are paid out after 1 month, 18 are paid out after 2 months, and 6 are paid out after 2 months.

  The jump-in demand information 163 is updated by the update unit 161b based on the demand information 162. FIG. 4 is a diagram for explaining the update of the dive demand information 163 by the update unit 161b. FIG. 4 shows the result of extracting the data related to the result of the dive demand of a certain type of product for one year from the demand information 162, and totaling the number of payouts for each registration month and every month until the payout date. The data related to the record of the jump-in demand is data in which “Y” is stored in the item of the jump-in flag and the date is stored in the item of the payout date. The registration month is the year and month stored in the registration date item. The number of months until the payout date is the difference between the year and month stored in the payout date item and the year when the demand is predicted.

  The update unit 161b calculates the average value of the number of payouts thus calculated for each registered month and for each number of months until the payout date, for each number of months until the payout date. Then, the updating unit 161b stores the calculated average value in the jump-in demand information 163 as a predicted value of jump-in demand generated during one month in association with the product type number. The timing and frequency at which the update unit 161b updates the dive demand information 163 may be arbitrarily determined. In the example illustrated in FIG. 4, the example in which the dive demand information 163 is updated using data for one year has been described. However, the period of data used to update the dive demand information 163 is a period for which demand is predicted. It may be changed according to the length.

  Based on the jump demand information 163 updated in this way, the demand prediction unit 161a calculates the jump demand Du of the specified type of product in the specified period using the following equation (2).

  Here, N is the length (number of months) of the specified period. di is a predicted value of what is paid out after i months out of the diving demand generated in one month, and is obtained from the diving demand information 163.

  For example, if the jump-in demand information 163 is as shown in FIG. 3 and the length of the specified period is three months, the demand forecasting unit 161a sets the jump-in demand Du for the type of product identified by “AAA” as follows: Calculate as follows.

Du = (23 × (3-0)) + (18 × (3-1)) + (6 × (3-2))
= (23 × 3) + (18 × 2) + (6 × 1)
= 285

  This formula shows that if the length of the specified period is 3 months, the dive demand to be paid out after 1 month is generated for 3 months, the dive demand to be paid out after 2 months is generated for 2 months, and after 3 months. It shows that the dive demand to be paid out is generated for one month.

  In this way, the demand prediction unit 161a calculates the jump-in demand by accumulating predicted values for each number of months until the payout date according to the length of the designated period. For this reason, the dive demand is accurately calculated.

  The coefficient information 164 holds information for predicting a demand that is assumed in advance. FIG. 5 is a diagram illustrating an example of the coefficient information 164. In the coefficient information 164, a coefficient indicating the accuracy of the inquiry is stored for each type of part, for each expectation, and for each number of months until the scheduled delivery date. The coefficient information 164 further stores the number of data of the demand information 162 used for calculating the coefficient as the number of samples for each part type and each expectation.

  For example, the data in the first row of the coefficient information 164 shown in FIG. 5 has the accuracy of what is registered in the demand information 162 with the expectation of “5” among the inquiries regarding the product of the type identified by “AAA”. 1.00 in the case of paying out in the current month (after 0 month), 0.97 in the case of paying out after 1 month, 0.21 in the case of paying out after 17 months, etc. ing.

  The coefficient information 164 is updated by the update unit 161b based on the demand information 162. 6 and 7 are diagrams for explaining the update of the coefficient information 164 by the updating unit 161b. FIG. 6 shows that data related to a certain type of product inquiries that have been registered as expectation level “5” and for which payment has been completed is extracted from demand information 162 for 18 months, and the number of months up to the date of payment and schedule The result of totaling the number of payouts for each number of months until delivery date is shown. The data relating to the inquiry for which payout has been completed is data in which “N” is stored in the item of the jump flag and the date is stored in the item of the payout date. The number of months until the scheduled delivery date is the difference between the year and month stored in the scheduled delivery date item and the year and month when the demand is predicted.

  The update unit 161b totals the number of payouts thus totaled for each number of months until the scheduled delivery date and for each number of months until the scheduled delivery date. In this way, the number of payouts (number of results) per month until the scheduled delivery date is calculated.

  Subsequently, the update unit 161b also obtains data related to the results of the inquiries related to the same type of product from the demand information 162, regardless of whether or not the payment is completed. Monthly data are extracted and aggregated for each number of months up to the payout date and every number of months up to the scheduled delivery date. The update unit 161b totals the number of payouts thus totaled for each number of months until the scheduled delivery date and for each number of months until the scheduled delivery date. Thus, the number of payouts per month until the scheduled delivery date is calculated.

  Then, the updating unit 161b calculates the accuracy of the inquiry by dividing the number of payouts by the planned number for each number of months until the scheduled delivery date. The updating unit 161b stores the accuracy for each number of months until the scheduled delivery date calculated in this manner as a coefficient in association with the value of “5” representing the product type number and the expected degree in the coefficient information 164.

  The update unit 161b calculates the accuracy for each number of months until the scheduled delivery date in the same manner for the expectation other than “5”, and stores the accuracy in the coefficient information 164 in association with the value indicating the expectation. The timing and frequency with which the updating unit 161b updates the coefficient information 164 may be arbitrarily determined.

  The updating unit 161b may store the accuracy calculated for each number of months until the scheduled delivery date in the coefficient information 164 as it is, but may store the accuracy in which variation is corrected by the moving average method in the coefficient information 164. As shown in the graph of FIG. 7, by calculating the moving average with a width of 3 months or 5 months, it is possible to obtain the accuracy with suppressed variation.

  In the example illustrated in FIG. 6, the example in which the coefficient information 164 is updated using data for 18 months has been described. However, the period of the data used to update the coefficient information 164 is the length of the period for which demand is predicted. You may change it according to.

  Based on the coefficient information 164 updated in this way, the demand prediction unit 161a is a value obtained by multiplying the expected number of products of a specified type in a specified period in advance by a coefficient. Calculated as an integrated value. Specifically, the demand prediction unit 161a calculates a demand De that is assumed in advance for a specified type of product in a specified period using the following equation (3).

  Here, M is an item of the scheduled delivery date, in which the identification number stored in the item of the type number in the data of the demand information 162 corresponding to the unpaid inquiry for the product matches the identification number of the specified type. Is the number of data types included in the specified period. Qj is a value stored in the scheduled number of items of j-th data that satisfies the above conditions. rj is a coefficient corresponding to the jth data that satisfies the above conditions. rj is acquired from the coefficient information 164 based on the identification number of the designated type, the number of months until the scheduled delivery date, and the value stored in the item of expectation of the jth data. The number of months until the scheduled delivery date is the number of months from the year and month stored in the registration date field of the jth data to the year and month stored in the scheduled delivery date field of the jth data.

  In this way, the demand prediction unit 161a calculates a presumed demand using a coefficient set for each number of months until the scheduled delivery date. As described above, the value of the item of expectation is set based on the judgment of the person in charge who negotiated with the customer, but the difference between the accuracy indicated by the item of expectation and the actual accuracy is limited to the delivery date. It may vary depending on the length of the period. For this reason, using the coefficient set for each number of months until the scheduled delivery date contributes to improvement in the accuracy of the calculated demand.

  For example, if the period until the delivery date is long, even if it is determined that the person in charge is likely to be able to pay, the payout may not be reached due to various circumstances, so the actual accuracy is higher than the accuracy indicated by the expectation item. May be low. Alternatively, if the period until the delivery date is short, the person in charge may be cautiously registering the accuracy with a low degree of accuracy, so the accuracy indicated by the item of expectation may be lower than the actual accuracy.

  FIG. 8 is a diagram illustrating an example of the coefficient pattern information 165. The coefficient pattern information 165 holds a plurality of coefficient combinations used for predicting a demand that is assumed in advance. In the coefficient pattern information 165 shown in FIG. 8, eight coefficient patterns each identified by the pattern numbers PT01 to PT08 are registered. The coefficient pattern is a combination of coefficients set for each degree of expectation.

  FIG. 9 is a diagram illustrating an example of the selection information 166. The selection information 166 holds the coefficient pattern selected from the coefficient patterns registered in the coefficient pattern information 165 for each part type and for each number of months (every three months) until the scheduled delivery date. The coefficient pattern information 165 and the selection information 166 are used instead of the coefficient information 164 when the number of data (the number of samples) of the demand information 162 used for calculating the coefficient set in the coefficient information 164 is small. The

  The selection information 166 is updated by the update unit 161b based on the demand information 162. 10 and 11 are diagrams for explaining the update of the selection information 166 by the update unit 161b. FIG. 10 shows all patterns obtained by combining the coefficient patterns registered in the coefficient pattern information 165 shown in FIG. 8, and the ratio between the demand predicted based on each pattern and the actual demand. Yes.

  For example, the data in the first row shown in FIG. 10 indicates that the number of months until the payout date is 0 to 2 months, 3 to 5 months, 6 to 8 months, 9 to 11 months, 12 to 14 months, and 15 to 17 months. The ratio calculated when the coefficient pattern identified by “PT01” is selected is shown. Further, for the type number AAA, “AAA-01”, “AAA-02” are added to 0 to 2 months, 3 to 5 months, 6 to 8 months, 9 to 11 months, 12 to 14 months, and 15 to 17 months, respectively. ”,“ AAA-03 ”,“ AAA-04 ”,“ AAA-05 ”,“ AAA-06 ”, and so on. (Only the type of type number is generated. In this example, since there are 14 types, alphabets are generated up to “NNN”), then a ratio of “1.21” is calculated for “AAA-01”. A ratio of “1.07” is calculated for “AAA-02”, and a ratio of “1.46” is calculated for “AAA-03”. The data on the first line further indicates that the ratio of “1.37” is calculated for the entire types identified by “AAA-01” to “NNN-06”, and that “AAA-01” to “NNN- The deviation of the ratio calculated for each type number between “06” is 0.16.

  For example, the actual demand of the type identified by “AAA-01” indicates that the data of the item of the type number among the data stored in the demand information 162 is “AAA-01” and the value of the item of withdrawal date Is calculated by extracting data for 18 months and adding the number of payouts corresponding to the data. The predicted value of the demand of the type identified by “AAA-01” is calculated by adding a value obtained by multiplying the value of the item of the planned number of extracted data by a coefficient. In order to obtain the coefficient, a pattern number corresponding to the number of months from the year and month at the time of data creation to the year and month of the payout date item is obtained from the combination of coefficient patterns of data in the same row. A coefficient pattern corresponding to the acquired pattern number is acquired from the coefficient pattern information 165, and a coefficient corresponding to the value of the item of expectation of the extracted data is acquired from the acquired coefficient pattern. Thus, a coefficient is acquired for each extracted data. The ratio is calculated by dividing the integrated value obtained by multiplying the predicted value by the coefficient by the actual demand integrated value.

  The update unit 161b calculates these values for all combinations of coefficient patterns, and then selects a combination in which the ratio of the entire products identified by the same type number is close to 1 and the deviation is small. Then, the update unit 161b stores the selected combination of coefficient patterns in the selection information 166. The ratio of the entire products identified by the same type number is close to 1 and the deviation is small. This means that the predicted demand is close to the actual demand regardless of the type number, that is, the demand is predicted. This means that the coefficient used for is suitable.

  The update unit 161b, for example, from the combinations in which the ratio of the entire products identified by the same type number is within the range of 0.98 to 1.02, as in the region R1 illustrated in FIG. The combination with the smallest deviation of the calculated ratio is selected. The range is set to an arbitrary width according to the distribution characteristics.

  As described above, the update unit 161 b selects a combination that is predicted to be close to the actual results from the combinations of coefficient patterns registered in the coefficient pattern information 165, and stores the selected combination in the selection information 166. In this method, since a coefficient pattern is prepared in advance, an extremely biased coefficient can be obtained even when the number of data (the number of samples) of the demand information 162 used for obtaining the actual demand is small. The nature is low.

  On the other hand, in the coefficient information 164 in which the coefficients are set recursively based on the actual results, if the number of samples is small, a biased coefficient may be obtained. For this reason, when the number of samples is small, the coefficient rj used in Equation (3) is acquired not from the coefficient information 164 but from the coefficient pattern information 165 and the selection information, thereby improving the demand prediction accuracy. it can.

  In this embodiment, the pattern number is set in the selection information 166 every three months instead of every month. By extending the period in this way, the number of samples used for calculating the ratio or the like when updating the selection information 166 can be increased, and the possibility that a more suitable coefficient pattern is selected increases.

  Next, a processing procedure of demand prediction processing executed by the demand prediction device 10 will be described with reference to FIG. FIG. 12 is a flowchart illustrating a processing procedure of the demand prediction process. The processing procedure illustrated in FIG. 12 is realized by the control unit 15 executing the demand prediction program 161. When executing the demand forecasting process, a period for forecasting demand and a type of product for forecasting demand are specified. The period and type may be specified in, for example, a command input via the input unit 12, specified in a request received via the communication unit 13, or stored in advance in the storage unit 16. It may be specified in a stored configuration file.

  As shown in FIG. 12, the control unit 15 uses the data of the demand information 162 corresponding to the inquiry that the value of the item of the type number matches the specified type number and the payout is scheduled within the specified period. Select (step S101).

 It is selected whether to perform a detailed examination of the expected number of demands acquired as inquiry information in advance or to calculate the number of jumps that are demands that are not assumed in advance (step S102).

  When selecting to perform a scrutiny of the expected number of demands assumed in advance, the control unit 15 sets the value of the item of the type number and the year set in the item of the scheduled delivery date from the year and month when the demand is predicted. Coefficients and sample numbers corresponding to the period up to the month and the value set in the expectation item are acquired from the coefficient information 164 (step S103). And the control part 15 compares the acquired number of samples with a threshold value (step S104). The threshold is set so that the coefficient held in the coefficient information 164 is guaranteed to have sufficient accuracy.

  When the number of samples is larger than the threshold (step S104, Yes), the control unit 15 multiplies the value of the scheduled number item by a coefficient (step S105) and adds the calculation result to the demand number (step S109). The number of demands is set to 0 at the start of the demand prediction process. Thereafter, the control unit 15 re-executes step S101 and subsequent steps until completion for all type numbers to be predicted (step S110).

  When the number of samples is not greater than the threshold (No at Step S104), the control unit 15 sets the value of the type number item and the item of the scheduled delivery date from the year and month when the demand is predicted from the selection information 166. A pattern number corresponding to a period up to a certain year and month is acquired (step S106). Subsequently, the control unit 15 acquires a coefficient corresponding to the value set in the item of expectation and the acquired pattern number from the coefficient pattern information 165 (step S107). Then, the control unit 15 multiplies the value of the scheduled number item by a coefficient (step S108), and adds the calculation result to the demand number (step S109). Thereafter, the control unit 15 re-executes step S101 and subsequent steps until completion for all type numbers to be predicted (step S110).

  When selecting to calculate the number of jumps in step S102, the control unit 15 calculates the jump-in demand based on the length of the specified period and the jump-in demand information 163 (step S112). Then, the control unit 15 adds the calculation result to the demand number (step S109). Thereafter, the control unit 15 re-executes step S101 and subsequent steps until completion for all type numbers to be predicted (step S110). The control unit 15 outputs the demand number as a prediction result (step S111), and ends the demand prediction process.

  As described above, according to the demand prediction device and the demand prediction method according to the present embodiment, demand can be predicted with high accuracy.

  In addition, the aspect of this invention shown by said Example can be arbitrarily changed in the range which does not deviate from the summary of this invention. For example, the program shown in the above embodiment may be divided into a plurality of modules or may be integrated with other programs. The function of the demand prediction device 10 may be appropriately distributed among a plurality of devices.

  In the above embodiment, an example is shown in which coefficients are set in units of months. However, the units for setting coefficients are not limited to this. For example, units of weeks, units of months, quarters, units of years, etc. But you can.

DESCRIPTION OF SYMBOLS 10 Demand prediction apparatus 11 Display part 12 Input part 13 Communication part 14 Medium reading part 15 Control part 16 Storage part 151 CPU
152 Memory 161 Demand Prediction Program 161a Demand Prediction Unit 161b Update Unit 162 Demand Information 163 Jump-in Demand Information 164 Coefficient Information 165 Coefficient Pattern Information 166 Selection Information

Claims (5)

A storage unit for storing demand information in which a scheduled delivery date of a product for which an inquiry has been made, a planned number to be paid out, and an expectation of an order are associated and held;
A demand prediction unit that predicts demand for the product by multiplying a value corresponding to the length of the period until the scheduled delivery date and the coefficient corresponding to the expectation by multiplying the planned number ;
The storage unit stores coefficient pattern information holding a plurality of coefficient patterns in which the coefficient is set for each expectation, and a value indicating which of the coefficient patterns is selected as a length of a period until the scheduled delivery date Further storing selection information to be held for each
The demand prediction unit acquires a value indicating which of the coefficient patterns is selected from the selection information based on a length of a period until the scheduled delivery date, and based on the acquired value and the expectation A demand prediction device that acquires the coefficient by which the planned number is multiplied from the coefficient pattern information . The storage unit further stores coefficient information that holds, as the coefficient, for each length of a period until the scheduled delivery date and for each expected degree, a probability that a product that has been inquired in the past is actually paid out,
The demand prediction device according to claim 1 , wherein the demand prediction unit obtains the coefficient by which the planned number is multiplied from the coefficient information when the number of products inquired in the past is larger than a threshold value. The demand prediction unit further predicts a dive demand based on the number of demands generated without inquiry for each predetermined period, and integrates a value obtained by multiplying the number by the coefficient. by adding the plunge demand, demand prediction apparatus according to claim 1 or 2, forecast demand. A demand prediction method executed by a demand prediction device,
Obtaining demand information that holds a scheduled delivery date of a product that has been inquired, a planned number to be paid out, and an expectation of an order,
Predicting the demand for the product by multiplying a value obtained by multiplying a length corresponding to the expected delivery date and a coefficient corresponding to the expectation by the expected number ,
In the step of acquiring the demand information, coefficient pattern information holding a plurality of coefficient patterns in which the coefficient is set for each expectation degree, and a value indicating which of the coefficient patterns is selected up to the scheduled delivery date And further acquiring selection information to be held for each length of period,
In the step of predicting the demand for the product, a value indicating which of the coefficient patterns is selected from the selection information is acquired based on a length of a period until the scheduled delivery date, and the acquired value and the expected degree And a demand prediction method for acquiring the coefficient by which the planned number is multiplied from the coefficient pattern information . Demand forecasting device
Obtaining demand information that holds a scheduled delivery date of a product that has been inquired, a planned number to be paid out, and an expectation of an order,
Predicting the demand for the product by multiplying a value obtained by multiplying the length corresponding to the expected delivery date and a coefficient corresponding to the expectation by the expected number ; and
In the step of acquiring the demand information, coefficient pattern information holding a plurality of coefficient patterns in which the coefficient is set for each expectation degree, and a value indicating which of the coefficient patterns is selected up to the scheduled delivery date And further acquiring selection information to be held for each length of period,
In the step of predicting the demand for the product, a value indicating which of the coefficient patterns is selected from the selection information is acquired based on a length of a period until the scheduled delivery date, and the acquired value and the expected degree Based on the above, a demand prediction program for acquiring the coefficient by which the planned number is multiplied from the coefficient pattern information .

Images