Prediction method and device
By obtaining sales information and correction parameters of old products in the same level, and combining sales and characteristic information on the reference date, the sales volume prediction of new products is optimized, and the problem of difficulty in predicting new products is solved, and more accurate sales prediction is achieved.
Patent Information
- Application Number
- CN202411246364.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-06
- Publication Date
- 2025-01-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
It is difficult for the existing technology to accurately predict the sales of new products, especially when the number of samples is limited and daily sales fluctuates a few days before the new products are launched.
By obtaining the sales information of old products of the same level in the new product period, determining the correction parameters, and using the correction parameters to correct the sales ratio of the new product, combining the sales volume and sales feature information of the reference date, sales volume prediction is carried out, and further optimization through fitting relationships Predict results.
It improves the accuracy and reliability of new product sales forecasts, reduces prediction deviations, and can predict new products more accurately after they are put on the shelves.
Smart Images
Figure CN119250883A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing, and in particular to a prediction method and device. Background Art
[0002] In the current sales forecast, after the goods are put on the shelves of the store, it is necessary to predict the future sales every day based on the existing sales data after the goods are put on the shelves, so as to determine the daily purchasing and allocation decisions.
[0003] However, it is often difficult to predict the sales of new products. This is because the number of samples of new products is limited in the first few days after the new products are launched, and it is difficult to obtain enough data to estimate future sales. In addition, the daily sales of new products usually fluctuate greatly, which increases the complexity of prediction.
[0004] Therefore, how to accurately predict the sales volume of commodities (especially new products) is a problem that needs to be solved urgently. Summary of the invention
[0005] The embodiments of the present invention provide a prediction method and device for accurately predicting the sales volume of commodities (especially new commodities).
[0006] In a first aspect, an embodiment of the present invention provides a prediction method, the method comprising: obtaining sales information of a first commodity, the sales information of the first commodity comprising a first sales ratio of the first commodity on each day from the Nth to the Mth day after being put on the shelves compared to the Nth day, N is a positive integer, and M is a positive integer greater than or equal to N; obtaining sales information of a second commodity, the sales information of the second commodity comprising a second sales ratio of each day from the Nth to the M+1th day after the second commodity is put on the shelves during the new product period compared to the Nth day, the second commodity being a commodity in the primary category to which the first commodity belongs, and Correction parameters are determined based on the first sales ratios of the Mth day compared to the Nth day and the second sales ratios of the Nth to Mth days after the second product is put on the shelves during the new product period compared to the Nth day. A reference sales ratio is determined based on the second sales ratio of the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day. The reference sales ratio is corrected using the correction parameters to obtain the predicted sales ratio of the first product on the M+1th day after it is put on the shelves compared to the Nth day. Based on the sales on the reference date and the predicted sales ratio, the sales of the first product on the M+1th day after it is put on the shelves are predicted, and the reference date is the day before the M+1th day.
[0007] Through the above scheme, the sales volume of the first commodity M+1 days can be predicted based on the known sales volume information of the first commodity, and the second commodity belonging to the same primary category as the first commodity is selected as a reference for sales volume prediction. Compared with the prior art that only uses the sales volume information of the first commodity to predict future sales volume, the sales volume prediction of the present invention is more accurate. The present invention uses the sales volume ratio of the second commodity to correct the sales volume ratio of the first commodity, so that the sales volume information of the first commodity can be corrected based on the known sales volume information of the first commodity and the historical data of commodities in the same category, so as to obtain sales volume information with more reference value.
[0008] Optionally, a correction parameter is determined based on the first sales ratios of the first product from the Nth to the Mth day after being put on the shelves compared to the Nth day and the second sales ratios of the second product from the Nth to the Mth day after being put on the shelves during the new product period compared to the Nth day, including: obtaining the first sales ratios of each day from the N+1th to the Mth day after the first product is put on the shelves compared to the Nth day, obtaining the second sales ratios of each day from the N+1th to the Mth day after the second product is put on the shelves during the new product period compared to the Nth day, calculating a first ratio of the first sales ratio of the first product from the Hth day after being put on the shelves compared to the Nth day and the second sales ratio of the second product from the Hth day after being put on the shelves during the new product period compared to the Nth day, where H is taken over each integer from N+1 to M, and the average of the first ratios corresponding to the N+1th day to the Nth day is taken as the correction parameter.
[0009] Optionally, before predicting the sales volume of the first product on the M+1 day after it is put on the shelves based on the ratio of the sales volume on the reference date to the predicted sales volume, the method further includes: obtaining the sales characteristic information of each day from the first to the Mth day after the first product is put on the shelves, and the sales characteristic information of the first product on the M+1 day after the first product is put on the shelves; taking a date from the first to the Mth day that is similar to the sales characteristic information of the M+1 day as a reference date, wherein the sales characteristic information includes one or more of the following: whether there is a VIP event, event discounts, whether it is a holiday, and the date from today.
[0010] Through the above scheme, the sales volume on the M+1 day can be predicted by the sales volume on the reference date. Since the reference date is similar to the sales feature information of the first product on the M+1 day, the sales volume on the reference date is more valuable for reference. On this basis, the sales volume of the first product can be predicted more accurately.
[0011] Optionally, based on the sales volume on the reference date and the predicted sales volume ratio, the sales volume of the first product on the M+1th day after the product is put on the shelves is predicted, including: taking the product of the sales volume on the reference date and the predicted sales volume ratio as the initial predicted sales volume on the M+1th day after the product is put on the shelves, fitting a first corresponding relationship based on the sales volume of each day from the first to the Mth day after the product is put on the shelves and the initial predicted sales volume, the first corresponding relationship is used to indicate the sales volume ratio of the first product on the Ith day after the product is put on the shelves compared to the Kth day, K is a positive integer, I is a positive integer greater than K and less than or equal to M, substituting the sales volume on the reference date into the first corresponding relationship to obtain the sales volume of the first product on the M+1th day after the product is put on the shelves. In this way, by fitting the first corresponding relationship, the prediction deviation is added, and the sales volume prediction is further optimized on the basis of the previous one.
[0012] Optionally, after fitting the first corresponding relationship and before substituting the sales volume on the reference date into the first corresponding relationship, the method further includes: fitting a second corresponding relationship for each product according to the sales volume of each product in the secondary category to which the first product belongs on each day from the first to the M+1th day after the new product is put on the shelves during the new product period, the second corresponding relationship being used to indicate the sales ratio of the first product on the Pth day after being put on the shelves compared to the Qth day, where Q is a positive integer, and P is a positive integer greater than Q and less than or equal to M, using the average value of each first parameter in the second corresponding relationship of each product in the secondary category to correct the first parameter in the first corresponding relationship, and correcting other parameters in the first corresponding relationship based on the corrected first parameter.
[0013] Through the above scheme, by using each product in the secondary category to correct the first corresponding relationship, a more accurate sales ratio and prediction deviation can be obtained. Compared with the prior art that only uses a single function for prediction, the present invention uses the historical secondary category products to correct the original function. While considering the actual sales situation of the first product, historical data is used as a reference to more accurately predict sales.
[0014] Optionally, the first product is a new product sold in the store, or the first product is a new product that the warehouse will provide to various stores. After predicting the sales volume of the first product on the M+1th day after it is put on the shelves, it also includes: providing the sales volume to various stores according to the allocation ratio of each store to the first product.
[0015] Through the above scheme, since the sales data of a single store is relatively small, selecting the sales data under the large warehouse can make up for the shortcomings of small and unstable sales data of a single store. Then allocating the sales to each store according to the proportion of the store can reduce the prediction deviation caused by the small sales data of new products.
[0016] In a second aspect, an embodiment of the present invention provides a prediction device, the prediction device comprising:
[0017] The acquisition module is used to obtain the sales information of the first product, the sales information of the first product includes the first sales ratio of each day from the Nth to the Mth day after the first product is put on the shelf compared to the Nth day, N is a positive integer, and M is a positive integer greater than N.
[0018] The acquisition module is also used to obtain the sales information of the second product, the sales information of the second product includes the second sales ratio of each day from the Nth to the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day, and the second product is a product in the first-level category to which the first product belongs.
[0019] The determination module is used to determine the correction parameters according to the first sales ratios of the first product from the Nth to the Mth day after being put on the shelves compared to the Nth day and the second sales ratios of the second product from the Nth to the Mth day after being put on the shelves during the new product period compared to the Nth day.
[0020] The determination module is also used to determine a reference sales ratio based on the second sales ratio of the second product on the M+1 day compared to the N day after the second product is put on the shelves during the new product period, and use a correction parameter to correct the reference sales ratio to obtain a predicted sales ratio of the first product on the M+1 day compared to the N day after the product is put on the shelves.
[0021] The prediction module is used to predict the sales volume of the first product on the M+1th day after it is put on the shelves according to the sales volume on the reference date and the predicted sales volume ratio, where the reference date is the day before the M+1th day.
[0022] In a possible implementation, the determination module is further used to: obtain each first sales ratio of the first product on each day from the N+1th to the Mth day after being put on the shelves compared to the Nth day, obtain each second sales ratio of the second product on each day from the N+1th to the Mth day after being put on the shelves during the new product period compared to the Nth day, calculate a first ratio of the first sales ratio of the first product on the Hth day after being put on the shelves compared to the Nth day and the second sales ratio of the second product on the Hth day after being put on the shelves during the new product period compared to the Nth day, H takes each integer from N+1 to M, and takes the average of the first ratios corresponding to the N+1th day to the Nth day as the correction parameter.
[0023] In a possible implementation, the acquisition module is further used to: obtain the sales characteristic information of the first product on each of the first to Mth days after it is put on the shelves, and the sales characteristic information that the first product will have on the M+1th day after it is put on the shelves; use the date from the first to Mth days that is similar to the sales characteristic information of the M+1th day as a reference date, wherein the sales characteristic information includes one or more of the following: whether there is a VIP event, event discounts, whether it is a holiday, and the date from today.
[0024] In a possible implementation, the determination module is further used to: predict the sales volume of the first product on the M+1th day after it is put on the shelves according to the sales volume on the reference date and the predicted sales volume ratio, including: taking the product of the sales volume on the reference date and the predicted sales volume ratio as the initial predicted sales volume of the first product on the M+1th day after it is put on the shelves, and fitting a first corresponding relationship according to the sales volume of each day from the first to the Mth day after the first product is put on the shelves and the initial predicted sales volume, the first corresponding relationship is used to indicate the sales volume ratio of the first product on the Ith day after it is put on the shelves compared to the Kth day, K is a positive integer, I is a positive integer greater than K and less than or equal to M, and the sales volume on the reference date is substituted into the first corresponding relationship to obtain the sales volume of the first product on the M+1th day after it is put on the shelves.
[0025] In a possible implementation, the determination module is also used to: fit a second correspondence relationship for each product based on the sales volume of each product in the secondary category to which the first product belongs from the first to the M+1th day after the new product is put on the shelves during the new product period, the second correspondence relationship being used to indicate the sales ratio of the first product on the Pth day compared to the Qth day after the product is put on the shelves, where Q is a positive integer, and P is a positive integer greater than Q and less than or equal to M, and use the average value of each first parameter in the second correspondence relationship of each product in the secondary category to correct the first parameter in the first correspondence relationship, and correct other parameters in the first correspondence relationship based on the corrected first parameter.
[0026] In one possible implementation, the first product is a new product sold in the store, or the first product is a new product that the warehouse will provide to various stores. The prediction module is also used to: predict the sales volume of the first product on the M+1th day after it is put on the shelves, and also includes: providing the sales volume to each store according to the allocation ratio of each store to the first product.
[0027] In a third aspect, the present application also provides a prediction device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method described in the various possible designs of the first aspect.
[0028] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, on which a computer program or instructions are stored. When the computer program or instructions are executed by a processor, the method described in the various possible designs of the first aspect is implemented.
[0029] In a fifth aspect, an embodiment of the present application further provides a computer program product, which, when executed on a computer, enables the computer to execute any of the methods described in the first aspect above.
[0030] These implementations or other implementations of the present application will be more concise and understandable in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0032] Figure 1 A schematic diagram of an interface for a merchant to list new products provided by an embodiment of the present invention;
[0033] Figure 2 A flowchart of a prediction method provided by an embodiment of the present invention;
[0034] Figure 3 A schematic diagram of the second sales ratio of an old product of a first-level category compared to the first day on each of the M+1 days of its new product period provided by an embodiment of the present invention;
[0035] Figure 4 A schematic diagram of a process for selecting a reference date provided by an embodiment of the present invention;
[0036] Figure 5 A schematic diagram of comparing the MAPE, accuracy, deviation rate and base_case rate of different solutions provided in an embodiment of the present invention;
[0037] Figure 6 A schematic diagram of a process for predicting the sales volume of a first product on the M+1th day after it is put on the shelf provided by an embodiment of the present invention;
[0038] Figure 7 A schematic diagram of a prediction device provided by an embodiment of the present invention;
[0039] Figure 8 A schematic diagram of another prediction device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0040] In order to make the purpose, technical solution and beneficial effects of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0041] Below, some terms in this application are explained. It should be noted that these explanations are for the convenience of understanding by those skilled in the art and do not constitute a limitation on the scope of protection claimed in this application.
[0042] 1. New product period.
[0043] The new product period is usually set according to the product category, shelf life and other attributes. For example, the new product period of short-shelf-life fruits is set to 14 days, and the new product period of long-shelf-life general merchandise is set to 21 days.
[0044] 2. Primary categories and secondary categories.
[0045] A primary category is a broader classification level in product classification, and a secondary category is a more specific level in product classification. For example, strawberries belong to fruits, and more specifically, to the berries category of fruits. In this case, "berries" is a secondary category, and "fruits" is a primary category.
[0046] 3. Linear regression.
[0047] Linear regression is one of the important prediction methods in the field of statistics and machine learning. Its core idea is to predict unknown data by fitting the linear relationship between independent variables (also called features, explanatory variables) and dependent variables (also called target variables, response variables).
[0048] When merchants sell goods, they will put some new products on the shelves. These new products will have separate categories, such as Figure 1 The interface diagram of a merchant listing new products is shown in the figure. It can be seen from the figure that there is a separate category of new products on the display interface, which can attract the attention of customers. However, as described in the background technology, it is difficult to predict the sales of new products. Currently, a preset single function (such as linear regression or prophet's logistic growth function) is used to predict the sales of new products. However, this method often cannot accurately predict the sales of new products and has a certain error. Therefore, how to accurately predict the sales of new products is a problem that needs to be solved urgently.
[0049] Based on this, the present invention provides a prediction method for accurately predicting the sales volume of a first commodity by referring to the sales volume information of a second commodity.
[0050] The present application scheme is described in detail below in conjunction with the accompanying drawings.
[0051] See also Figure 2 , a flow chart of a prediction method is shown, which specifically includes the following steps:
[0052] Step 210: Obtain sales information of the first product.
[0053] The first product may be a new product, that is, a product that is put on the shelf for the first time and has just been put on the shelf. The sales information of the first product includes the first sales ratio of each day from the Nth to the Mth day after the first product is put on the shelf compared to the Nth day, where N is a positive integer and M is a positive integer greater than or equal to N.
[0054] For example, the first product has been sold for 4 days, and the sales volume of the first 4 days is 1, 2, 3, and 4 respectively. Assuming that the sales volume of the 5th day is predicted based on the sales volume information of the 1st to 4th day, then N is 1, M is 4, and the sales volume information of the first product is:
[0055] Step 220: Obtain sales information of the second product.
[0056] The second product is a product in the first-level category to which the first product belongs, and the sales information of the second product includes the second sales ratio of each day from the Nth to the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day.
[0057] Optionally, the second product is each product in the first-level category to which the first product belongs, and the average sales ratio of each product in the first-level category to which the first product belongs on each day from the Nth to the M+1th day after the new product is put on the shelves during the new product period compared to the Nth day is taken as the second sales ratio. For example, the first product is strawberry, and as introduced in the above-mentioned explanation of terms, the first-level category of strawberry is fruit, then the second product is a fruit group, that is, each product in the fruit category, such as bananas, apples, blueberries, etc. Calculate the sales ratio of each product in the first-level category to which the first product belongs (such as bananas, apples, blueberries, etc.) on each day from the Nth to the M+1th day after the new product is put on the shelves during the new product period compared to the Nth day, and take the average sales ratio of each product as the second sales ratio.
[0058] For example, Table 1 shows the second sales ratio of the old products in the first-level category on each day of the M+1th day of its new product period compared to the first day. Figure 3 This is the line graph corresponding to Table 1. It can be seen from the figure that when a new product is first launched, the sales trend first increases and then tends to be flat.
[0059]
[0060]
[0061] Table 1
[0062] You can select the sales information of old products in the first-level category during a certain period of history. For example, if the new product is strawberry, you can find old products (i.e., fruit group) in the first-level commodity category during a certain period of history (for example, within 90 days) and obtain the sales information of the old products during their new product period.
[0063] Since the second product is an old product that has been on the shelves for a long time, we need to select the data of the second product during its new product period. Assume that the sales volume of the second product in the first 5 days after it was put on the shelves is 2, 3, 4, 5, and 6 respectively. Based on the sales volume from the 1st day to the 5th day, the sales volume information of the second product is:
[0064] Step 230: Determine correction parameters according to the first sales ratios of the first product from the Nth to the Mth day after being put on the shelves compared to the Nth day and the second sales ratios of the second product from the Nth to the Mth day after being put on the shelves during the new product period compared to the Nth day.
[0065] For example, the first sales ratios of the first commodity calculated above are 1, 2, 3, and 4 respectively; the second sales ratios of the second commodity are 1, 1.5, 2, and 2.5 respectively; then the correction parameters are determined based on the first sales ratio and the second sales ratio.
[0066] Optionally, the method for determining the correction parameter may be: obtaining each first sales ratio of the first product on each day from the N+1th to the Mth day after being put on the shelves compared to the Nth day, obtaining each second sales ratio of the second product on each day from the N+1th to the Mth day after being put on the shelves during the new product period compared to the Nth day, calculating a first ratio of the first sales ratio of the first product on the Hth day after being put on the shelves compared to the Nth day and the second sales ratio of the second product on the Hth day after being put on the shelves during the new product period compared to the Nth day, H being each integer from N+1 to M, and the average value of each first ratio corresponding to the N+1th day to the Nth day is taken as the correction parameter.
[0067] For example, divide the sales ratio of the first product on the second day by the sales ratio of the second product on the second day to get the first ratio value on the second day: Divide the sales ratio of the first product on the third day by the sales ratio of the second product on the third day to get the sales ratio value on the third day: Divide the sales ratio of the first product on the fourth day by the sales ratio of the second product on the fourth day to get the sales ratio value for the fourth day: Then take the average value of the first ratio as the correction parameter. The complete process of calculating the correction parameter is as follows:
[0068]
[0069] That is, the calculated correction parameter is 1.48.
[0070] Step 240: Determine a reference sales ratio value according to the second sales ratio of the second product on the M+1th day compared to the Nth day after the second product is put on the shelves during the new product period.
[0071] Continuing with the above example, in step 220, it has been obtained that the sales ratio of the second product on the 5th day after the new product is put on the shelves compared to the 1st day is 3, so the reference sales ratio is 3.
[0072] Step 250: Use the correction parameter to correct the reference sales ratio to obtain the predicted sales ratio of the first product on the M+1th day after it is put on the shelves compared to the Nth day.
[0073] Optionally, using the correction parameter to correct the reference sales ratio may be: multiplying the correction parameter by the reference sales ratio to obtain the predicted sales ratio. For example, if the reference sales ratio calculated above is 3 and the correction parameter is 1.48, then the predicted sales ratio is 3*1.48=4.44.
[0074] Step 260: Based on the sales volume on the reference date and the predicted sales volume ratio, predict the sales volume of the first product on the M+1th day after it is put on the shelves.
[0075] Through the above steps 210 to 260, the sales volume of the first commodity in M+1 days can be predicted based on the known sales information of the first commodity, and the second commodity belonging to the same primary category as the first commodity is selected as a reference for sales prediction. Compared with the prior art that only uses the sales information of the first commodity to predict future sales, the sales prediction of the present invention is more accurate. The present invention uses the sales ratio of the second commodity to correct the sales ratio of the first commodity, so that the sales information of the first commodity can be corrected based on the known sales information of the first commodity and based on the historical data of commodities in the same category, so as to obtain sales information with more reference value.
[0076] In the above step 260, the reference date is the day before M+1 days. The reference date can be selected based on historical experience or sales characteristics, for example, Figure 4 A flow chart of selecting a reference date is shown, which specifically includes the following steps:
[0077] Step 410: Obtain the sales characteristic information of the first product on each of the first to Mth days after the product is put on the shelves, and the sales characteristic information that the first product will have on the M+1th day after the product is put on the shelves.
[0078] In sales forecasting, some important factors need to be considered, such as event discounts, holidays, etc. These important factors need to be considered when looking for a reference date. In addition, the closer the reference date is to today, the higher the value of the reference.
[0079] Optionally, the selling feature information may be: whether there is a VIP activity, activity discount (divided into intervals of 0.05, 0.1, 0.2, 0.3, 0.4, 0.5), whether it is a holiday, and the time from now (number of weeks).
[0080] Based on experience, you can sort the above important factors by priority and choose a suitable priority scheme to find the reference date. Alternatively, you can set different schemes, and select the one with better prediction results by counting the prediction results of different schemes. For example, the following four schemes are set to find the reference date, and the priority of the important factors in each scheme decreases from the front to the back:
[0081] Solution 1: whether there are VIP activities, event discounts, whether it is a holiday, and the time from now;
[0082] Solution 2: whether there are VIP activities, event discounts, time from now, whether it is a holiday;
[0083] Solution 3: Whether there are VIP activities, event discounts, and time from now (i.e. holidays and non-holidays have the same priority);
[0084] Plan 4: Is it a VIP event, event discount, or is it a holiday (i.e., if the target date is a holiday, holiday sales will be prioritized).
[0085] Predict the sales volume of the first commodity according to the above four schemes respectively, compare the prediction results of the above four schemes, comprehensively select the scheme with the best prediction effect, and use the best scheme to find the reference date.
[0086] Optionally, compare the prediction results of the schemes, such as mean absolute percentage error (MAPE), accuracy, bias, adverse cases (i.e., cases with large prediction bias), etc. See Figure 5 The diagram shown is a schematic diagram comparing the MAPE, accuracy, deviation rate and base_case rate of the above schemes. After comparison, scheme 3 with small MAPE, high accuracy, small deviation rate and few adverse conditions is finally selected as the scheme for selecting the reference date.
[0087] Step 420: Use the date from the first to the Mth day that has similar sales characteristic information to the M+1th day as a reference date.
[0088] For example, find the date with the same VIP activity as the M+1 day, the same activity discount as the M+1 day, and the closest date to the present as the reference date. Then, based on the sales volume on the reference date and the predicted sales volume ratio, predict the sales volume of the first product on the M+1 day after it is put on the shelves.
[0089] Through the above steps 410 and 420, the sales volume on the M+1 day can be predicted by the sales volume on the reference date. Since the reference date is similar to the sales characteristic information of the first commodity on the M+1 day, the sales volume on the reference date is more valuable for reference. On this basis, the sales volume of the first commodity can be predicted more accurately.
[0090] Optionally, before predicting the sales volume of the first product on the M+1th day after it is put on the shelves based on the ratio of the sales volume on the reference date to the predicted sales volume, the sales volume on the reference date needs to be adjusted. For example, the sales volume on the reference date can be corrected based on the discount on the reference date.
[0091] Since the sales feature information of M+1 day and the reference date is different, such as different activities or holidays, the sales volume of the reference date needs to be adjusted. For example, if you want to predict the sales volume on November 7, there is a 50% off event on the reference date November 1, and the sales volume is higher, while there is no discount on November 7. At this time, it is necessary to correct the sales volume on November 1 to an estimated value without a discount, and then consider the growth ratio. There are many ways to correct sales volume. For example, you can use machine learning methods to learn the elasticity coefficient to correct sales volume, or use the discount ratio to correct sales volume, etc., which are not limited here. This embodiment adopts a method of correcting sales volume based on the discount ratio, and the formula for sales volume correction is as follows:
[0092] Sale_restore=Sale_reference*discount_1
[0093] Where Sale_restore is the revised sales volume on the reference date, Sale_reference is the sales volume before the revision on the reference date, and discount_1 is the discount on the reference date.
[0094] Assuming that there is a 50% off promotion on the reference date November 1st, and the sales volume is 100, then the revised sales volume will be 0.5 times the original sales volume, that is, the revised sales volume will be 100*0.5=50.
[0095] Through the above scheme, the sales volume on the reference date can be corrected. Correcting the sales volume on the reference date can reduce the prediction error caused by the difference in sales characteristic information between the reference date and the M+1 day.
[0096] It is understandable that the above steps of correcting the sales volume on the reference date and selecting the reference date (i.e., step 410 and step 420) are not in any order. In other words, all sales volume after the first product is put on the shelf can be corrected first, and then a reference date close to the sales characteristic information of the M+1 day can be selected; or a reference date close to the sales characteristic information of the M+1 day can be selected first, and then the sales volume on the reference date can be corrected.
[0097] Optionally, in step 260, based on the sales volume on the reference date and the predicted sales volume ratio, the formula for predicting the sales volume of the first product on the M+1th day after it is put on the shelves is as follows:
[0098]
[0099] Wherein, Sale_forecast is the sales volume predicted for the first product on day M+1, k is the predicted sales volume ratio, which can be calculated by the above step 250, and discount_2 is the discount predicted for the first product on day M+1.
[0100] Continuing with the above example, assuming that there is also a 50% discount event on November 7, when the predicted sales ratio is 1, the predicted sales on November 7 is: That is, the predicted sales on November 7th and the actual sales on November 1st are the same.
[0101] However, if November 7 is not a 50% discount event, but a 20% discount event, that is, in the above formula, discount_2 is 0.8, and there is no reference date with a 20% discount in history, it is difficult to predict how much sales will be achieved during the 20% discount event on November 7. In other words, the sales volume and discount on November 7 are not in an ideal linear relationship. Based on this, it is necessary to try to find a date that is close to the sales feature information of November 7 as a reference date (that is, the above steps 310 and 320), so that the prediction error caused by different discounts can be avoided and the accuracy of the prediction can be improved.
[0102] Optionally, when the discounts on the reference date and the M+1 day are the same, the formula for predicting the sales volume of the first product on the M+1 day after it is put on the shelves based on the sales volume on the reference date and the predicted sales volume ratio in step 260 is as follows:
[0103] Sale_forecast=Sale_restore*k
[0104] For example, continuing with the above example, assuming that the sales volume after correction on the reference date is 1, after the predicted sales volume ratio k is calculated to be 4.44 in step 250, the predicted sales volume of the first product is calculated to be 1*4.44=4.44 in step 260.
[0105] Optionally, the sales volume of the first product on the M+1th day after it is put on the shelf can also be predicted using the following formula:
[0106] Sale_forecast = Sale restore *k*m
[0107] Among them, m is the holiday coefficient. That is, based on the previous formula, this formula also takes into account the impact of holidays on the sales volume of the first product.
[0108] Optionally, in the above step 260, the sales volume of the first product on the M+1 day after being put on the shelves can also be predicted by adding the sales volume deviation, that is, based on the sales volume on the reference date, the predicted sales volume ratio and the sales volume deviation, the sales volume of the first product on the M+1 day after being put on the shelves is predicted. Figure 6 A flow chart of predicting the sales volume of a first product on the M+1th day after it is put on the shelf is shown, which specifically includes the following steps:
[0109] Step 610: The product of the sales volume on the reference date and the predicted sales volume ratio is used as the initial predicted sales volume of the first product on the M+1th day after it is put on the shelves.
[0110] This step is the same as the above step 260 and will not be repeated here.
[0111] Step 620: According to the sales volume of the first product on each of the first to Mth days after the product is put on the shelves and the initial predicted sales volume, a first corresponding relationship is obtained by fitting.
[0112] Among them, the first corresponding relationship is used to indicate the sales ratio of the first product on the Ith day after being put on the shelf compared to the Kth day, K is a positive integer, and I is a positive integer greater than K and less than or equal to M.
[0113] Optionally, a linear regression formula can be used to fit the first corresponding relationship, where the first corresponding relationship is: y=k*x+b; where x is the sales volume of the first product on the Kth day after it is put on the shelf, y is the sales volume of the first product on the Ith day after it is put on the shelf, k is the sales volume ratio of the first product on the Ith day after it is put on the shelf to the Kth day, and b is the sales volume deviation. The calculation method of the first corresponding relationship y=k*x+b is as follows:
[0114]
[0115] in, is the mean value of x, is the mean value of y.
[0116] For example, assuming that the sales volume of the first product on the first to seventh days after it is put on the shelves is 1, 2, 3, 4, 5, 7, 9, and the initial predicted sales volume (i.e., the sales volume on the eighth day) is 11, then the sales volume on the Ith day and the Kth day are selected as x and y respectively. Assuming that the selected data are (1, 3), (2, 5), (3, 7), (4, 9), (5, 11), the average value of the data x and the average value of the data y can be calculated according to the above formula (1) and formula (2), which are respectively:
[0117]
[0118] Then, the parameters k and b in the first corresponding relationship y=k*x+b are calculated according to formula (3) and formula (4). The calculation process is as follows:
[0119]
[0120] b=7-2*3=1
[0121] Finally, the first fitting relationship is y=2x+1.
[0122] Step 630: According to the sales volume of each product in the secondary category to which the first product belongs on each day from the first day to the M+1th day after the new product is put on the shelves during the new product period, a second corresponding relationship for each product is obtained by fitting.
[0123] Among them, the second corresponding relationship is used to indicate the sales ratio of the first product on the Pth day after being put on the shelves compared to the Qth day, Q is a positive integer, and P is a positive integer greater than Q and less than or equal to M.
[0124] Optionally, a linear regression formula may be used to fit the second corresponding relationship. The specific fitting process may refer to step 620 and will not be described in detail here.
[0125] Step 640: Determine the average value of each first parameter in the second corresponding relationship of each commodity in the secondary category to which the first commodity belongs.
[0126] After fitting multiple second corresponding relationships in step 630, multiple k can be obtained, and then the average value of k is taken. Specifically, the following second corresponding relationship can be obtained in step 630:
[0127] y m+1 =k m+1 *x m+1 +b m+1 ……(1)
[0128] y m+2 =k m+2 *x m+2 +b m+2 ……(2)
[0129] …
[0130] y m+i =k m+i *x m+i +b m+i …(i)
[0131] Since there are many products in the secondary category to which the first product belongs, for example, the secondary category of strawberries is berries, and the secondary category of nuts includes grapes, blueberries, mulberries, etc., by calculating the second correspondence of each product in the secondary category to which the first product belongs, multiple second correspondences can be obtained, each of which contains the first parameter k m , take k m The average value of
[0132] It is understandable that the number of secondary category products selected in this step (i.e., the value of i in the above formula) can be determined based on actual conditions or obtained based on experience, and is not limited here.
[0133] Step 650: Use the average value to correct the first parameter in the first corresponding relationship.
[0134] Using step 640 calculated The k calculated in step 620 can be corrected in many ways, such as addition or multiplication, which is not limited here. The present invention uses multiplication to correct the first parameter in the first corresponding relationship, that is, the corrected first parameter k1 is:
[0135]
[0136] Step 660: Correct other parameters in the first corresponding relationship according to the corrected first parameter.
[0137] Substitute the corrected parameter k1 into k in the first correspondence y=k*x+b, and recalculate b according to formula (4) in the above step 620. In this way, the parameters k and b in the first correspondence y=k*x+b calculated in step 620 are corrected again to obtain a new corrected first correspondence.
[0138] Step: 670: Substitute the sales volume of the reference date into the first corresponding relationship to obtain the sales volume of the first product on the M+1th day after it is put on the shelf.
[0139] Finally, the sales volume on the reference date is substituted into the corrected first corresponding relationship to obtain the sales volume of the first product on the M+1th day after it is put on the shelf.
[0140] Through the above steps, the sales volume of the first product on day M+1 can be further predicted more accurately. In the above steps 610 and 620, by fitting the first corresponding relationship, the prediction deviation is added, and the sales volume prediction is further optimized on the basis of the previous one; in the above steps 630 to 660, the first corresponding relationship is corrected using each product in the secondary category, and a more accurate sales volume ratio and prediction deviation can be obtained. Compared with the prior art that only uses a single function for prediction, the present invention uses the historical secondary category products to correct the original function, and while considering the actual sales of the first product, the historical data is used as a reference, so that the sales volume can be predicted more accurately.
[0141] Optionally, the first product in the present invention may be a new product sold by a store, or may be a new product that a large warehouse will provide to each store. If the first product is a new product that a large warehouse will provide to each store, then after the above step 260, the process further includes: providing the sales volume to each store according to the allocation ratio of each store to the first product.
[0142] Since the sales data of a single store is relatively small, selecting the sales data under the large warehouse can make up for the shortcomings of small and unstable sales data of a single store. Then allocating the sales to each store according to the proportion of the store can reduce the prediction deviation caused by the small sales data of new products.
[0143] Optionally, the sales volume of new products under the large warehouse needs to meet the following three conditions:
[0144] 1. The average daily sales volume of the new product is less than the first preset value, that is, the first product is a low-selling product.
[0145] 2. The number of stores under the large warehouse is greater than the second preset value. If the number of stores is small, the prediction method is the same as the prediction method for a single store, which does not reflect the advantage of using a large warehouse for sales forecasting. In addition, if the number of stores is small, it may also be an abnormal situation. Therefore, the number of stores needs to be greater than the second preset value.
[0146] 3. The first store that sells each product is the same. This is to prevent new products from being put on the shelves in one store today and in another store tomorrow, which would affect the accuracy of the data.
[0147] The above introduces a method for predicting the sales volume of the first product. The following three specific examples illustrate the above method. It should be noted that in the above steps 310 to 360, an example is used to introduce a method for predicting the sales volume on the fifth day of listing by referring to the first day of listing of the first product. The following examples 1 and 2 continue to use the data in the above steps 310 to 360 for example, and example 3 illustrates steps 610 to 670.
[0148] Example 1: Predict the sales volume of the first product on the fifth day of listing based on the second day of listing.
[0149] Step a: Get the sales information of the first product.
[0150] The sales volume of the first product from the 2nd day to the 4th day are 2, 3, and 4 respectively. (That is, N is 2 and M is 4 in the above step 210), then the sales volume information of the first product is:
[0151] Step b: Obtain sales information of the second product.
[0152] When the second product was first put on the shelves, the sales volume from the second day to the fifth day was 4, 5, and 6 respectively. The sales volume information of the second product is:
[0153] Step c: Determine correction parameters according to the first sales ratios of the first product on the 2nd to 4th days after being put on the shelves compared to the 2nd day and the second sales ratios of the second product on the 2nd to 4th days after being put on the shelves during the new product period compared to the 2nd day.
[0154] Divide the sales ratio of the first product on the third day by the sales ratio of the second product on the third day to get the sales ratio value on the third day: Divide the sales ratio of the first product on the fourth day by the sales ratio of the second product on the fourth day to get the sales ratio value for the fourth day: Then take the average value of the first ratio as the correction parameter. The complete process of calculating the correction parameter is as follows:
[0155]
[0156] Step d: Determine a reference sales ratio value according to the second sales ratio of the second product on the fifth day after the product is put on the shelves during the new product period compared to the second day.
[0157] The reference sales ratio obtained from step b is 2.
[0158] Step e: Use the correction parameter to correct the reference sales ratio to obtain the predicted sales ratio of the first product on the 5th day after it is put on the shelves compared with the 2nd day.
[0159] The predicted sales ratio is 2*1.16=2.32.
[0160] Step f: Based on the sales volume on the reference date and the predicted sales volume ratio, predict the sales volume of the first product on the fifth day after it is put on the shelves.
[0161] Assuming that the sales volume on the reference date is 2, the sales volume of the first product on the fifth day after it is put on the shelves is: 2*2.32=4.64.
[0162] Example 2: Predict the sales volume of the first product on the 5th day based on the 4th day of its launch.
[0163] Step a: Get the sales information of the first product.
[0164] The sales volume of the first product on the 4th day is 4. (That is, N is 4 and M is 4 in the above step 210), then the sales volume information of the first product is:
[0165] Step b: Obtain sales information of the second product.
[0166] When the second product was first put on the shelves, the sales volume on the 4th and 5th days were 5 and 6 respectively. The sales volume information of the second product is:
[0167] Step c: Determine the correction parameters.
[0168] Since the first product was only on the shelves for 4 days, there is no other reference data to calculate the correction parameter, so the correction parameter here is 1.
[0169] Step d: Determine a reference sales ratio value according to the second sales ratio of the second product on the fifth day compared to the fourth day after the second product is put on the shelves during the new product period.
[0170] The reference sales ratio obtained from step b is 1.2.
[0171] Step e: Use the correction parameter to correct the reference sales ratio to obtain the predicted sales ratio of the first product on the fifth day after it is put on the shelves compared to the fourth day.
[0172] The predicted sales ratio is 1*1.2=1.2.
[0173] Step f: Based on the sales volume on the reference date and the predicted sales volume ratio, predict the sales volume of the first product on the fifth day after it is put on the shelves.
[0174] Assuming that the sales volume on the reference date is 4, the sales volume of the first product on the fifth day after it is put on the shelves is: 4*1.2=4.8.
[0175] Example 3: Based on the sales volume on the reference date, the predicted sales volume ratio and the sales volume deviation, predict the sales volume of the first product on the M+1th day after it is put on the shelves. Assuming that the first product is put on the shelves on the 19th, the sales volume on the 25th (the seventh day on the shelves) is predicted on the 27th (the ninth day on the shelves). The reference date is the 22nd (the fourth day on the shelves).
[0176] Step a: The product of the sales volume on the 22nd and the predicted sales volume ratio is taken as the initial predicted sales volume of the first product on the 27th.
[0177] This step can refer to the above example 1 and example 2, and will not be repeated here.
[0178] Step b: Use the linear regression formula to fit the first corresponding relationship of the first product.
[0179] This step can refer to the above step 620, which will not be described again. Assume that the first corresponding relationship obtained by fitting is y=0.941x+0.877, where x is the sales volume of the first product on the 22nd, and y is the sales volume of the first product on the 27th.
[0180] Step c: fitting to obtain a second corresponding relationship of the secondary category products, and determining an average value of each first parameter in the second corresponding relationship.
[0181] Assume that the growth rates on the 23rd (5th day on the shelf) and the 24th (6th day on the shelf) compared to the 22nd are 1.927 and 0.542 respectively. Then the average value is
[0182] Step d: using the average value to correct the first parameter in the first corresponding relationship.
[0183] The corrected first parameter is 0.941*1.23=1.157.
[0184] Step e: Correct other parameters in the first corresponding relationship according to the corrected first parameter.
[0185] According to the formula (3) in the above step 620, the corrected b=0.068 is calculated, and finally the first corresponding relationship is y=1.157x+0.068.
[0186] Step f: Substitute the sales volume on the 22nd into the first corresponding relationship to obtain the sales volume of the first product on the 27th day after it is put on the shelf.
[0187] Based on the same concept, an embodiment of the present application also provides a prediction device, which can execute the prediction method introduced in the above content.
[0188] See also Figure 7 , a schematic diagram of the structure of a prediction device provided in an embodiment of the present application is given, such as Figure 7 As shown, the prediction device comprises:
[0189] The acquisition module 701 is used to obtain the sales information of the first product, and the sales information of the first product includes the first sales ratio of each day from the Nth to the Mth day after the first product is put on the shelf compared to the Nth day, where N is a positive integer and M is a positive integer greater than N.
[0190] The acquisition module 701 is also used to obtain the sales information of the second product, and the sales information of the second product includes the second sales ratio of each day from the Nth to the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day. The second product is a product in the first-level category to which the first product belongs.
[0191] The determination module 702 is used to determine the correction parameters according to the first sales ratios of the first product from the Nth to the Mth day after being put on the shelves compared to the Nth day and the second sales ratios of the second product from the Nth to the Mth day after being put on the shelves during the new product period compared to the Nth day.
[0192] Determination module 702 is also used to determine a reference sales ratio based on the second sales ratio of the second product on the M+1 day compared to the N day after the second product is put on the shelves during the new product period, and use the correction parameter to correct the reference sales ratio to obtain the predicted sales ratio of the first product on the M+1 day compared to the N day after the product is put on the shelves.
[0193] The prediction module 703 is used to predict the sales volume of the first product on the M+1th day after it is put on the shelves according to the sales volume on the reference date and the predicted sales volume ratio, where the reference date is the day before the M+1th day.
[0194] In a possible implementation, the determination module 702 is further used to: obtain each first sales ratio of the first product on each day from the N+1th to the Mth day after being put on the shelves compared to the Nth day, obtain each second sales ratio of the second product on each day from the N+1th to the Mth day after being put on the shelves during the new product period compared to the Nth day, calculate a first ratio of the first sales ratio of the first product on the Hth day after being put on the shelves compared to the Nth day and the second sales ratio of the second product on the Hth day after being put on the shelves during the new product period compared to the Nth day, H takes each integer from N+1 to M, and takes the average of each first ratio corresponding to the N+1th day to the Nth day as the correction parameter.
[0195] In a possible implementation, the acquisition module 701 is further used to: obtain the sales characteristic information of the first product on each of the first to Mth days after it is put on the shelves, and the sales characteristic information that the first product will have on the M+1th day after it is put on the shelves; use the date from the first to Mth days that is similar to the sales characteristic information of the M+1th day as a reference date, wherein the sales characteristic information includes one or more of the following contents: whether there is a VIP event, event discounts, whether it is a holiday, and the date from today.
[0196] In a possible implementation, the determination module 702 is further used to: predict the sales volume of the first product on the M+1th day after it is put on the shelves according to the sales volume on the reference date and the predicted sales volume ratio, including: taking the product of the sales volume on the reference date and the predicted sales volume ratio as the initial predicted sales volume of the first product on the M+1th day after it is put on the shelves, and fitting a first corresponding relationship according to the sales volume of each day from the first to the Mth day after the first product is put on the shelves and the initial predicted sales volume, the first corresponding relationship is used to indicate the sales volume ratio of the first product on the Ith day after it is put on the shelves compared to the Kth day, K is a positive integer, I is a positive integer greater than K and less than or equal to M, and the sales volume on the reference date is substituted into the first corresponding relationship to obtain the sales volume of the first product on the M+1th day after it is put on the shelves.
[0197] In a possible implementation, the determination module 702 is further used to fit a second correspondence relationship for each product based on the sales volume of each product in the secondary category to which the first product belongs from the first to the M+1th day after the new product is put on the shelves during the new product period, wherein the second correspondence relationship is used to indicate the sales ratio of the first product on the Pth day after the product is put on the shelves compared to the Qth day, where Q is a positive integer, and P is a positive integer greater than Q and less than or equal to M. The first parameter in the first correspondence relationship is corrected using the average value of each first parameter in the second correspondence relationship of each product in the secondary category, and the other parameters in the first correspondence relationship are corrected based on the corrected first parameter.
[0198] In one possible implementation, the first product is a new product sold in the store, or the first product is a new product that the warehouse will provide to various stores. The prediction module 703 is also used to: predict the sales volume of the first product on the M+1th day after it is put on the shelf, and also includes: providing the sales volume to each store according to the allocation ratio of each store to the first product.
[0199] See also Figure 8 , showing a schematic diagram of the structure of another prediction device provided in an embodiment of the present application, such as Figure 8 As shown, the test device includes: a memory 801 and a processor 802, and the processor 802 is coupled to the memory 801. The memory 801 is used to store program instructions, and the processor 802 is used to call the program instructions stored in the memory 801 and execute the above prediction method according to the obtained program.
[0200] Optionally, the data processing device may further include an interface circuit 803, which may be a transceiver or an input / output interface. The input / output interface is used to input and / or output information, where output may be understood as sending and input may be understood as receiving. The processor 802 may communicate with other devices in the test device or other devices other than the test device through the interface circuit 803 to obtain information required for executing the above data processing method.
[0201] When the data processing device 800 is used to implement Figure 2 When the method is shown, the processor 802 is used to implement the functions of the above-mentioned acquisition module 701, determination module 702, and prediction module 703.
[0202] It is understandable that the processor in the embodiments of the present application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.
[0203] The memory in the embodiments of the present application can be a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, a register, a hard disk, a mobile hard disk, a compact disc read-only memory (CD-ROM) or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. The storage medium can also be a component of the processor.
[0204] Based on the same technical concept, an embodiment of the present invention also provides a computer-readable storage medium, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, the computer executes the above-mentioned transaction processing method.
[0205] Based on the same technical concept, an embodiment of the present invention further provides a computer-readable program product, which, when executed, enables a computer to execute the above transaction processing method.
[0206] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0207] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0208] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0209] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0210] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.
Claims
1. A prediction method, characterized in that: The method comprises: Acquire sales information of a first product, where the sales information of the first product includes a first sales ratio of each day from the Nth to the Mth day after the first product is put on the shelves compared to the Nth day, where N is a positive integer and M is a positive integer greater than or equal to N; Acquire sales information of a second product, where the sales information of the second product includes a second sales ratio of each day from the Nth day to the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day, and the second product is a product in the first-level category to which the first product belongs; Determine the correction parameter according to each first sales ratio of the first product from the Nth to the Mth day after the product is put on the shelves compared to the Nth day and each second sales ratio of the second product from the Nth to the Mth day after the product is put on the shelves during the new product period compared to the Nth day; Determine a reference sales ratio according to the second sales ratio of the second product on the M+1 day after the product is put on the shelves in the new product period compared to the N day, and use the correction parameter to correct the reference sales ratio to obtain a predicted sales ratio of the first product on the M+1 day after the product is put on the shelves compared to the N day; According to the sales volume on the reference date and the predicted sales volume ratio, the sales volume of the first product on the M+1th day after being put on the shelves is predicted, and the reference date is the day before the M+1th day.
2. The method according to claim 1, characterized in that The determining of the correction parameter according to each first sales ratio of the first product from the Nth to the Mth day after the product is put on the shelves compared to the Nth day and each second sales ratio of the second product from the Nth to the Mth day after the product is put on the shelves during the new product period compared to the Nth day includes: Obtain a first sales ratio of the first product on each of the N+1th to Mth days after the product is put on the shelves compared to the Nth day; Obtain the second sales ratio of each day from the N+1th day to the Mth day after the second product is put on the shelves during the new product period compared to the Nth day; Calculate a first ratio of a first sales ratio of the first product on the Hth day after it is put on the shelves compared to the Nth day and a second sales ratio of the second product on the Hth day after it is put on the shelves during the new product period compared to the Nth day, where H is an integer from N+1 to M; The average value of the first ratios corresponding to the N+1th day to the Nth day is used as the correction parameter.
3. The method according to claim 1, characterized in that Before predicting the sales volume of the first product on the M+1th day after the product is put on the shelves based on the sales volume on the reference date and the predicted sales volume ratio, the method further includes: Acquire the sales characteristic information of each day from the first day to the Mth day after the first product is put on the shelves, and the sales characteristic information that the first product will have on the M+1th day after the first product is put on the shelves; Using the date with similar sales characteristic information to the M+1th day from the first to the Mth day as the reference date; The sales characteristic information includes one or more of the following: whether there is a VIP event, event discounts, whether it is a holiday, and the date from today.
4. The method according to claim 1, characterized in that The step of predicting the sales volume of the first product on the M+1th day after the product is put on the shelves based on the sales volume on the reference date and the predicted sales volume ratio includes: The product of the sales volume on the reference date and the predicted sales volume ratio is used as the initial predicted sales volume of the first product on the M+1th day after the product is put on the shelves; A first corresponding relationship is obtained by fitting according to the sales volume of each day from the first to the Mth day after the first product is put on the shelves and the initial predicted sales volume, where the first corresponding relationship is used to indicate the sales volume ratio of the first product on the Ith day after the product is put on the shelves to the Kth day, where K is a positive integer and I is a positive integer greater than K and less than or equal to M; The sales volume on the reference date is substituted into the first corresponding relationship to obtain the sales volume of the first product on the M+1th day after it is put on the shelf.
5. The method according to claim 4, characterized in that After the first corresponding relationship is obtained by fitting, and before the sales volume on the reference date is substituted into the first corresponding relationship, the method further includes: According to the sales volume of each product in the secondary category to which the first product belongs on each day from the first to the M+1th day after the new product is put on the shelves during the new product period, a second corresponding relationship of each product is fitted, where the second corresponding relationship is used to indicate the sales volume ratio of the first product on the Pth day after the product is put on the shelves to the Qth day, where Q is a positive integer and P is a positive integer greater than Q and less than or equal to M; The first parameter in the first corresponding relationship is corrected using the average value of each first parameter in the second corresponding relationship of each commodity in the secondary category, and the other parameters in the first corresponding relationship are corrected according to the corrected first parameter.
6. The method according to claim 1, characterized in that The first product is a new product sold in the store; or The first product is a new product that the warehouse will provide to each store. The predicted sales volume of the first product on the M+1th day after the product is put on the shelves also includes: The sales volume is provided to each store according to the allocation ratio of each store to the first commodity.
7. A prediction device, characterized in that: The prediction device comprises: an acquisition module, configured to acquire sales information of a first product, wherein the sales information of the first product includes a first sales ratio of each day from the Nth to the Mth day after the first product is put on the shelves compared to the Nth day, where N is a positive integer and M is a positive integer greater than N; The acquisition module is further used to acquire sales information of a second product, the sales information of the second product including a second sales ratio of each day from the Nth to the M+1th day after the second product is put on the shelves during the new product period compared to the Nth day, the second product being a product in the first-level category to which the first product belongs; a determination module, configured to determine a correction parameter according to each first sales ratio of the first product from the Nth to the Mth day after the product is put on the shelves compared to the Nth day and each second sales ratio of the second product from the Nth to the Mth day after the product is put on the shelves during the new product period compared to the Nth day; The determination module is further configured to determine a reference sales ratio value based on a second sales ratio of the second product on the M+1th day compared to the Nth day after the second product is put on the shelves during the new product period, and to correct the reference sales ratio value using the correction parameter to obtain a predicted sales ratio value of the first product on the M+1th day compared to the Nth day after the first product is put on the shelves; The prediction module is used to predict the sales volume of the first product on the M+1th day after it is put on the shelves according to the sales volume on the reference date and the predicted sales volume ratio, where the reference date is the day before the M+1th day.
8. A prediction device, characterized in that: include: A processor, wherein the processor is coupled to a memory, the memory is used to store a computer program or an instruction, and the processor is used to execute the computer program or the instruction to implement the method according to any one of claims 1 to 6.
9. A computer-readable storage medium, characterized in that: It stores a computer program executable by a computer device, and when the program is run on the computer device, the computer device executes the steps of any method described in claims 1 to 6.
10. A computer program product, characterized in that When the method is executed on a computer, the computer is enabled to execute the method according to any one of claims 1 to 6.