JPH08115369A - Sales amount prediction system - Google Patents

Abstract

PURPOSE: To realize a precise in-month daily sales amount prediction in which periodic variation and week variation in a month are both reflected by using a predicted value of a daily sales amount constitution ratio on a prediction object day in the month and a predicted value of a sales amount in the month including the prediction object day. CONSTITUTION: An arithmetic unit 4 consists of an arithmetic function 6 which performs the four rules of arithmetic operations and various function operations and a sales amount predicting function 7 which predicts the daily sales amounts of prediction object articles. A predicted value of a daily sales amount condition ratio on the prediction object day in the month is calculated from the predicted values of the predicted daily sales amount constitution rate and week sales amount constitution rate in the month including the prediction object day, and a predicted value of the sales amount on the prediction object day is calculated from the calculated predicted value and the predicted value of the sales amount in the month including the prediction object day calculated from monthly sales amount actual result information. Consequently, the precise in-month daily sales amount prediction in which week variation and variation by the days of the week in the month are both reflected is actualized without reference to the prediction skillfulness of a planner.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to demand forecasting, which is carried out as a pre-stage operation such as sales planning and production planning, and in particular, it is often performed on a weekly basis by day and day. The present invention relates to a sales volume prediction method suitable for sales volume estimation.

[0002]

2. Description of the Related Art Prediction target units in demand forecast differ depending on the nature of a plan that utilizes the forecast result. In the demand forecast in the large schedule plan or the medium schedule plan, the forecast target is often a monthly unit. In addition, as the data used for the demand forecast, the sales volume results are often used in terms of the ease of obtaining the data and the reliability of the data itself. Therefore, in the demand forecast of a product, the actual sales amount of the product is used as the forecast data, and the time series forecasting technique or the multiple regression forecasting technique known as a known technique can be directly applied as a concrete forecasting method. However, in the execution plan for the next month such as a small schedule plan, the forecast target is a daily unit, so the above-mentioned known technology that models the demand fluctuation characteristics by a mathematical statistical approach is used as the daily demand forecast method. It is difficult to apply directly. As a publicly known example aimed at solving this, a "demand forecasting device" (Japanese Patent Application No. 3-13
7701 specification). In this known example, a seasonal index showing the ratio of each month to the planned amount and a seasonal index showing the ratio of each season to the demand forecast amount of each month are set in advance,
The forecasted demand amount for each month and the forecasted demand amount for each season are calculated based on these indexes, and the forecasted demand amount for each day is calculated based on the forecast demand amount for each season.

[0003]

In the above prior art, the daily forecast demand amount is calculated without considering the day-of-the-week factor, which is an important fluctuation factor of the daily demand, so that the day-to-day detailed management is required. However, there was a problem that accurate forecasting results could not be expected in the demand forecasting at the time of formulating the execution plan for the following month. Furthermore, in the conventional technology, there is a problem that the seasonal index or the seasonal index, which is the input information of the prediction calculation, cannot be accurately set unless the planner has the predictive power to gain an accurate insight into the current state of the market and future trends. It was

The object of the present invention is to predict the sales volume for realizing accurate daily sales volume forecast within the month, which reflects both weekly fluctuations within the month and day-of-week fluctuations, without being influenced by the skill of the planner. To provide a method.

[0005]

In order to achieve the above object, the present invention provides an input device, an output device, a storage device including information on a daily sales volume result of a product and a storage area, and an internal storage device. In the sales volume forecasting method using a computer having a computing device for calculating the daily sales volume of the product on the forecast target date input from the input device based on the information of Based on the weekly ranking information by day within the month of the day, the daily sales volume results are aggregated by weekly ranking to create sales volume performance information by weekly ranking by month, and the sales volume performance information by weekly ranking by month is created. Monthly sales volume results are aggregated by month to create monthly sales volume performance information, and the sales volume composition ratio is calculated on a weekday basis based on the monthly sales volume performance information by week and the daily sales volume performance information. However, the sales volume performance information by month and by month Calculate the weekly sales volume composition ratio within the month from the sales volume performance information by rank, and calculate the forecast value of the weekday sales volume composition ratio within the month including the forecast target day from the above-mentioned weekday sales volume composition ratio The forecast value of the weekly sales volume composition ratio within the month including the forecast target day is calculated from the weekly sales volume composition ratio, and the prediction target is calculated from the forecast value of the weekday sales volume composition ratio and the forecast value of the weekly sales volume composition ratio. The forecast value of the daily sales volume composition ratio within the month of the day is calculated, and the forecast value of the sales volume within the month including the forecast target day is calculated from the monthly sales volume performance information to predict the sales volume within the month. The predicted value of the sales volume on the prediction target day is calculated from the value and the predicted value of the daily sales volume composition ratio, and the result is output to the output device.

[0006]

By providing the above means, the following operational effects can be obtained.

[0007] The sales volume composition ratio on a weekly day basis is calculated by aggregating the daily sales volume results by weekly ranking based on the weekly ranking information for each day within each month of the year including the forecast target day. It is calculated from the weekly sales amount actual result information and the daily sales amount actual result, and the predicted value of the day-to-day sales amount composition ratio within the month including the forecast target day is calculated from the calculated weekly sales amount composition ratio. In addition, the weekly sales composition ratio within the month is calculated by aggregating the weekly sales volume results of the monthly weekly sales volume information by month and creating the monthly sales volume actual information and the monthly weekly sales volume actual information by month. From the calculated monthly weekly sales composition ratio, the predicted value of the weekly sales volume composition ratio within the month including the forecast target day is calculated.

From the predicted values of the day-to-day sales volume composition ratio and the weekly sales volume composition ratio within the month including the prediction target day calculated by the above means, the predicted value of the daily sales volume composition ratio within the month of the prediction target day is calculated. Then, the predicted value of the sales amount on the prediction target day is calculated from the calculated predicted value and the predicted value of the sales amount within the month including the prediction target date calculated from the monthly sales amount actual information. As a result, it is possible to realize an accurate daily sales forecast within a month that reflects both weekly fluctuations within the month and day-of-week fluctuations, without being influenced by the skill of the planner.

[0009]

1 is a block diagram of an apparatus according to the present embodiment, which comprises an input device 1, an output device 2, a storage device 3, an arithmetic unit 4, and a control unit 5 for controlling them.

As shown in FIG. 2, the storage device 3 has a sales record information table 200 having a prediction target product name storage field 201 and a sales volume performance data field 202 showing the daily sales volume performance of the prediction target product. As shown in FIG. 3, a prediction target date information table 300 having a year storage column 301, a month storage column 302, a day storage column 303, and a day of the week storage column 304 input from the input device 1 is stored. .

The arithmetic unit 4 comprises an arithmetic function 6 for performing four arithmetic operations and various function operations, and a sales volume forecasting function 7 for forecasting the daily sales volume of a forecast target product according to the present invention.

The processing operation and information transmission operation of the sales volume forecasting function 7, which is the core of the present invention, will be described below with reference to the flowchart of FIG.

<Step 401> Creation of weekly ranking table for each day within month Based on the first month-by-month 500 for one year including the forecast target date set by the man-machine interface shown in FIG. A week-by-day weekly ranking table 600 (FIG. 6) that defines which week of the month each day corresponds to is created. Specifically, as shown in the flowchart of FIG.
First, after setting the target month for determining the weekly ranking by day from the monthly start date 500 (step 701), the days from the first month of the target month to the day before the first month of the next month are set earlier. The day-by-day weekly ranking is determined every 7 days in order from the other party, and the result is registered in the daily weekly ranking table 600 within the month (step 702). If the weekly ranking by day of the month has been determined for all months of the year, this process is terminated; otherwise, the process returns to step 701 (step 703).

<Step 402> Aggregation of daily sales volume by weekly ranking Based on the daily weekly ranking table 600 within the month, daily sales volume actuals up to the forecast target date are aggregated by weekly ranking. Specifically, as shown in the flowchart of FIG. 8, first, based on the weekly ranking by day of the month in the weekly ranking table by day of the month 600,
The actual daily sales volume of the forecast target day before the previous year is aggregated by weekly ranking by month (step 801). Further, the daily sales volume results from the beginning of the year including the forecast target date to the month in which the monthly sales volume results are fixed are totaled by the weekly ranking by month (step 802). Finally, the totalized results of steps 801 and 802 are combined and registered in the monthly sales volume by week ranking table 900 shown in FIG. 9 (step 803).

<Step 403> Aggregation of sales volume results by week ranking by month Based on the sales volume performance table 900 by weekly ranking by month, the month in which the sales volume results for the current month including the forecast target date are confirmed. The weekly sales volume results up to are aggregated by month and registered in the monthly sales volume performance table 1000 shown in FIG.

<Step 404> Calculation of Sales Volume Composition Ratio on Weekday Weekly Sales Volume by Month and Weekly Sales Volume and Sales Performance Information Table 20
Corresponding to the daily sales volume registered in 0, calculate the daily sales volume composition ratio in the monthly weekly sales volume,
It is registered with the day-of-week day-by-day sales volume composition ratio table 1100 shown in FIG.

<Step 405> Calculation of weekly sales volume composition ratio within month Monthly sales volume registered in monthly sales volume performance table 1000 and sales volume performance table 900 by month ranking by month
Corresponding to the sales volume by weekly ranking by month registered in, calculate the sales volume composition ratio by weekly ranking within the month,
The weekly sales volume composition ratio table 12 for each month shown in FIG.
Register with 00.

<Step 406> Calculation of predicted value of day-to-day sales volume composition ratio by week ranking within month including prediction target day Based on table 1100 of weekly sales volume composition day by week ranking by month Within month including prediction target day Calculate the forecast value of the day-to-day sales volume composition ratio of. A specific processing procedure will be described with reference to the flowchart shown in FIG.

[Step 1301] Import of Prediction Target Date The month M of the prediction target date is imported from the prediction target date information table 300.

[Step 1302] Month-by-week week-by-week ranking by day of week Sales volume composition ratio information retrieval Month-by-week week-by-week ranking Day-by-day sales volume composition-by-day table 1100 From month-to-month forecast of the target N of month M The day-to-day sales volume composition ratio is taken out and registered in the day-of-week sales volume composition ratio prediction data table 1400 shown in FIG. In addition, the reason why the sales volume composition ratio by day of week by week for the latest N months of the month M of the forecast target day is extracted is to incorporate the latest market trend into the sales volume forecast.

[Step 1303] Calculation of average value of sales volume composition ratio for each day of the week Based on the data table 1400 for forecasting the sales volume composition ratio of each day, the average of the sales volume composition ratio is calculated for each day by the formula shown in Formula 1. Calculate the value.

[0022]

[Equation 1]

[Step 1304] Calculating predicted value of day-to-day sales volume composition ratio The average value of day-to-day sales volume composition ratio calculated in step 1303 is standardized by the calculation formula of Formula 2, and is standardized within the month including the forecast target day. It is registered in the day-of-week sales volume composition ratio table 1500 shown in FIG. 15 as the predicted value of the weekly sales volume composition ratio.

[0024]

[Equation 2]

<Step 407> Calculation of forecast value of weekly sales volume composition ratio within month including forecast target day Based on weekly sales volume composition ratio table 1200 by month, weekly sales volume composition ratio within month including forecast target date Calculate the predicted value of. A specific processing procedure will be described with reference to the flowchart shown in FIG.

[Step 1601] Import of Prediction Target Date The month M of the prediction target date is imported from the prediction target date information table 300.

[Step 1602] Import of weekly sales volume composition ratio information by month From the weekly sales volume composition ratio table 1200 by month for the latest G months of the month M of the forecast target date, the month composition of the same number of week configurations as the month M Fig. 17 shows the weekly sales volume composition ratio of
The data is registered in the weekly sales volume composition ratio prediction data table 1700 shown in FIG. It should be noted that the reason why the weekly sales volume composition ratio by month for the latest G months of the month M of the forecast target day is taken out is to incorporate the latest market trend into the sales volume forecast.

[Step 1603] Calculation of average value of sales volume composition ratio for each week Based on the weekly sales volume composition ratio prediction data table 1700, the average value of the sales volume composition ratio is calculated for each week by the formula 3 calculate.

[0029]

(Equation 3)

[Step 1604] Calculating weekly sales volume composition ratio predictive value The average value of the weekly sales volume composition ratio calculated in step 1603 is standardized by the formula (4), and it is calculated within the month including the forecast target date. As the predicted value of the weekly sales volume composition ratio, it is registered in the weekly sales volume composition ratio table 1800 shown in FIG.

[0031]

[Equation 4]

<Step 408> Calculation of predicted value of monthly sales volume composition ratio on forecast target day Day of week sales volume composition ratio in day of week sales volume composition table 1500 and weekly sales volume in weekly sales volume composition ratio table 1800 Based on the composition ratio, the predicted value Z of the monthly sales volume composition ratio is calculated by the equation (5) and registered in the prediction target daily sales volume composition ratio table 1900 shown in FIG.

[0033]

(Equation 5)

<Step 409> Prediction of Monthly Sales Volume Including Prediction Target Date Based on the monthly sales volume results in the monthly sales volume performance table 1000, forecast of monthly sales volume including the prediction target date The value is calculated, and the forecast monthly sales volume table 2000 shown in FIG.
Register with. ARIMA (Auto Regressive Integral), which is a typical technique for time series forecast, is used for monthly sales forecast.
A well-established prediction technique such as the ted Moving Average) model may be applied. Therefore, since this processing is within the category of easy analogy, detailed description thereof is omitted here.

<Step 410> Calculation of predicted value of sales volume on forecast target day Sales volume composition ratio Z of forecast target day sales volume in forecast target day sales volume composition table 1900 and forecast monthly sales volume table 20
Based on the predicted monthly sales volume F in 00, the predicted value E of the sales volume on the prediction target day is calculated by Equation 6, and registered in the prediction target day sales volume table 2100 shown in FIG.

[0036]

[Equation 6] E = F * Z (Equation 6) <Step 411> Output of prediction result The contents of the forecast target day sales volume 2100 are output to the output device 2
Output to. An example of the output is shown in FIG. In the figure, the forecast sales volume on days other than the forecast target date (March 15, 1994) shows the result of executing the processing of steps 401 to 410 for each individual day.

By the processing steps as described above, it is possible to accurately predict the daily sales volume within the month, which reflects both the weekly fluctuation and the day-of-week fluctuation.

[0038]

According to the present invention, it is possible to realize a highly accurate daily sales forecast within a month, which reflects both weekly fluctuations within the month and day-of-week fluctuations, without being influenced by the skill of the planner. This can be expected to improve the accuracy of the execution plan for the next month, which is often managed on a weekly and day-by-day basis.

[Brief description of drawings]

FIG. 1 is a block diagram of an apparatus according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of a sales performance information table related to prediction target products.

FIG. 3 is an explanatory diagram of a prediction target date information table.

FIG. 4 is an explanatory diagram of a processing operation of a sales volume prediction method and an information transmission operation.

FIG. 5 is an explanatory diagram showing an example of a man-machine interface for inputting the first month / month by month including the prediction target date.

FIG. 6 is an explanatory diagram of a weekly daily weekly ranking table.

FIG. 7 is a flowchart showing a procedure for creating a weekly daily ranking table within a month.

FIG. 8 is a flowchart showing a procedure for creating a sales volume performance table by month and weekly ranking.

FIG. 9 is an explanatory diagram of a sales amount actual result table according to weekly ranking according to month.

FIG. 10 is an explanatory diagram of a monthly sales volume performance table.

FIG. 11 is an explanatory diagram of a day-to-day sales volume composition ratio table for each month and week.

FIG. 12 is an explanatory diagram of a monthly sales volume composition ratio table for each month.

FIG. 13 is a flowchart showing a procedure for calculating a predicted value of a day-to-day sales volume composition ratio for each weekly ranking within a month including a prediction target day.

FIG. 14 is an explanatory diagram of a data table for day-to-day sales volume composition ratio prediction.

FIG. 15 is an explanatory diagram of a day-to-day sales volume composition ratio table.

FIG. 16 is a flowchart showing a procedure for calculating a predicted value of a monthly sales volume composition ratio within a month including a prediction target date.

FIG. 17 is an explanatory diagram of a weekly sales volume composition ratio prediction data table.

FIG. 18 is an explanatory diagram of a weekly sales volume composition ratio table.

FIG. 19 is an explanatory diagram of a forecast target daily sales volume composition ratio table.

FIG. 20 is an explanatory diagram of a predicted monthly sales volume table.

FIG. 21 is an explanatory diagram of a forecast target day sales volume table.

FIG. 22 is an explanatory diagram showing an example of output of sales volume prediction results.

[Explanation of symbols]

404 ... Calculation of sales volume composition ratio on weekday of the week, 405 ... Calculation of weekly sales volume composition ratio of month, 406 ... Calculating predicted value of daytime sales volume composition ratio for each week within month including prediction target day, 407 ... Prediction target Calculating the forecast value of the weekly sales volume composition ratio including the day, 4
08: Calculation of the forecast value of the monthly sales volume composition ratio on the forecast target day.

Claims (4)

[Claims] 1. An input device, an output device, a storage device including information on a daily sales volume result of a product and a storage area, and
In a sales volume forecasting method using a computer having a computing device for calculating the daily sales volume of a product on the forecast target date input from the input device based on the information in the storage device, including the forecast target date Based on weekly daily ranking information within each month of each year, the daily sales volume results are aggregated by weekly rankings to create monthly weekly sales volume actuals information, and the monthly sales by weekly rankings are performed. Monthly sales volume results of volume performance information are aggregated by month to create monthly sales volume performance information, and sales volume is calculated on a weekly basis based on the monthly sales volume performance information by month and the daily sales volume performance information by month. Calculate the composition ratio, calculate the weekly sales volume composition ratio within the month from the monthly sales volume performance information and the monthly weekly ranking sales volume performance information, and calculate the forecast target day from the sales volume composition ratio within the weekday. Calculate the forecast value of the day-to-day sales volume composition within the included month and The forecast value of the weekly sales volume composition ratio within the month including the forecast target day is calculated from the weekly sales volume composition ratio, and the forecast target day is calculated from the forecast value of the weekday sales volume composition ratio and the forecast value of the weekly sales volume composition ratio. Calculate the forecast value of the daily sales volume composition ratio within the month, calculate the forecast value of the monthly sales volume including the forecast target date from the monthly sales volume actual information, and forecast the sales volume within the month And a predicted value of the daily sales volume composition ratio, and calculates a predicted value of the sales volume on the prediction target day, and outputs the result to the output device. 2. The week-by-day weekly ranking information within each month of each day including the forecast target day according to claim 1, wherein only the first month and day of each month entered by the input device is the week of each day within the month. Sales volume forecasting method that generates rankings. 3. The calculation of the forecast value of the day-to-day sales volume composition ratio in the month including the forecast target day according to claim 1, based on the actual value of the sales volume composition ratio of the nearest weekday of the forecast target day. A sales volume forecasting method that calculates the average value of the sales volume composition ratio for each day and standardizes the average value of the sales volume composition ratio for each day so that the total of them is 100%. 4. In claim 1, the calculation of the forecast value of the daily sales volume composition ratio within the month of the forecast target day is performed by the weekly sales volume composition ratio of the month having the same number of week configurations as the month including the forecast target day. A sales volume forecasting method that calculates the average value of the weekly sales volume composition ratio for each weekly ranking based on the values and normalizes the average value of the weekly sales volume composition ratio so that the total is 100%. .