CN116862574A

CN116862574A - Product demand prediction method, device, electronic equipment and readable storage medium

Info

Publication number: CN116862574A
Application number: CN202210303008.XA
Authority: CN
Inventors: 王全; 张丹阳; 王晶; 陈秋丽; 李京圆; 寇天天; 樊歆羽
Original assignee: SF Technology Co Ltd
Current assignee: SF Technology Co Ltd
Priority date: 2022-03-24
Filing date: 2022-03-24
Publication date: 2023-10-10

Abstract

On one hand, the method selects product sales data based on the product sales characteristic value of a target product, and because the product sales characteristic value contains the change relation of the product sales along with the sales time, the method can be used for judging whether the historical sales data can correctly embody the market demand condition of the current target product, and further selects effective product sales data from the historical sales data when the historical sales data contains redundancy data which cannot be correctly embodied, compared with the method for predicting by directly adopting the historical sales data, the demand of the target product is more accurate, and the quantity of dead stock can be effectively reduced according to the product characteristics of the automobile industry. On the other hand, the method does not depend on training of sample data, so that a large amount of sample data is not required to be prepared, and the prediction cost of product requirements is reduced.

Description

Product demand prediction method, device, electronic equipment and readable storage medium

Technical Field

The application relates to the technical field of express logistics, in particular to a product demand prediction method and device, electronic equipment and a readable storage medium.

Background

In the field of inventory, in order to make up products in real time, inventory backlog is not caused by excessive replenishment, products are usually predicted to determine whether replenishment is needed or not, and the quantity of replenishment is what.

With the development of the automobile industry, facing the current situation that the stock management improves the profit margin when the number of long-tail commodities in the automobile after-market sku (Stock Keeping Unit, minimum stock unit) is large, it is important to accurately predict the demand of the sku in the automobile industry to effectively improve the profit margin.

However, the current method for predicting the sku demand in the automobile assembly industry is based on the experience judgment of the staff, so that the accuracy of the prediction is poor.

Disclosure of Invention

The application provides a product demand prediction method, a device, electronic equipment and a readable storage medium, and aims to solve the problem that the accuracy of prediction is poor because the sky demand prediction method in the existing automobile distribution industry is based on experience judgment of staff.

In a first aspect, the present application provides a method for product demand prediction, comprising:

acquiring historical sales data of a target product to be predicted;

determining a product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time;

Selecting product sales data from the historical sales data according to the product sales characteristic value;

and predicting the demand of the target product according to the product sales data.

In one possible implementation manner of the present application, the determining the characteristic value of the sales volume of the target product according to the sales time in the historical sales data and the sales volume corresponding to each sales time includes:

determining the average value of the sales intervals of the target products according to the sales time of each product;

according to the sales volume of each product, determining a historical sales volume average value and a historical sales volume change average value of the target product;

and determining the historical sales volume average value, the historical sales volume change average value and the sales interval average value as product sales volume characteristic values of the target product.

In one possible implementation manner of the present application, the selecting product sales data from the historical sales data according to the product sales feature value includes:

if the product sales characteristic value meets a preset first condition, counting the sales time of each product in the historical sales data to obtain the ex-warehouse frequency of the target product;

If the ex-warehouse frequency is smaller than a preset frequency threshold, dividing the sales time in the historical sales data into preset time periods, counting the total product sales in each preset time period, and taking the ex-warehouse time period and the corresponding product sales which are not zero as the product sales data;

if the ex-warehouse frequency is greater than or equal to a preset frequency threshold, comparing each sales time in the historical sales data with a preset time condition to obtain a target sales time meeting the preset time condition, and taking the target sales time and the corresponding product sales volume as product sales data.

In one possible implementation manner of the present application, the predicting the demand of the target product according to the product sales data includes:

determining a target sales volume average value and a target sales volume change average value of the target product according to each product sales volume in the product sales data;

determining a target service level corresponding to the target product according to the product sales characteristic value;

determining a safety stock value of the target product according to the target service level and the target sales volume change mean value;

And predicting the demand of the target product according to the safety stock value and the target sales average value.

In one possible implementation manner of the present application, after the predicting the demand of the target product according to the product sales data, the method further includes:

if the product sales characteristic value does not meet a preset first condition, acquiring a target inventory value of the target product;

and if the target stock value is smaller than or equal to the safety stock value of the target product, replenishing according to the demand, wherein the safety stock value is obtained according to the preset service level corresponding to the product sales characteristic value and the product sales data.

counting the historical sales volume ratio of the target product in each preset warehouse;

dividing the demand according to the historical sales volume occupation ratio of each preset warehouse to obtain the replenishment volume of each preset warehouse;

and replenishing the goods in each preset warehouse according to the replenishing quantity of each preset warehouse.

In one possible implementation manner of the present application, the obtaining historical sales data of the target product to be predicted includes:

acquiring sales data of a plurality of alternative products;

determining the income amount of each alternative product according to the preset product cost and the preset sales amount corresponding to each alternative product;

and screening each candidate product according to the profit amount of each candidate product to obtain a target product with the corresponding profit amount larger than a preset value threshold, and taking the sales data of the target product as historical sales data.

In a second aspect, the present application provides a product demand prediction apparatus comprising:

the acquisition unit is used for acquiring historical sales data of the target product to be predicted;

the determining unit is used for determining the product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time;

a selection unit, configured to select product sales data from the historical sales data according to the product sales feature value;

and the prediction unit is used for predicting the demand of the target product according to the product sales data.

In a possible implementation of the application, the determining unit is further configured to:

In a possible implementation of the application, the selection unit is further configured to:

In a possible implementation of the application, the prediction unit is further configured to:

In a possible implementation of the application, the acquisition unit is further configured to:

acquiring sales data of a plurality of alternative products;

In a third aspect, the present application also provides an electronic device, the electronic device comprising a processor, a memory and a computer program stored in the memory and executable on the processor, the processor executing the steps of any one of the product demand prediction methods provided by the present application when the processor invokes the computer program in the memory.

In a fourth aspect, the present application also provides a readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of any of the product demand forecasting methods provided by the present application.

In summary, the product demand prediction method provided by the application comprises the following steps: acquiring historical sales data of a target product to be predicted; determining a product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time; selecting product sales data from the historical sales data according to the product sales characteristic value; and predicting the demand of the target product according to the product sales data. On one hand, the product demand prediction method provided by the application selects the product sales data based on the product sales characteristic value of the target product, and because the product sales characteristic value contains the change relation of the product sales along with the sales time, the product demand prediction method can be used for judging whether the historical sales data can correctly reflect the market demand condition of the current target product, and further selecting effective product sales data from the historical sales data when the historical sales data contains redundant data which cannot be correctly reflected, compared with the method for predicting by directly adopting the historical sales data, the obtained demand of the target product is more accurate, and the quantity of dead stock can be effectively reduced according to the product characteristics of the automobile industry. On the other hand, the product demand prediction method provided by the application does not depend on the training of sample data, so that compared with a method for predicting by adopting machine learning, the prediction error generated by the fact that the sample data is not representative can be avoided, and the predicted data base is the historical sales data of the target product, so that a large amount of sample data is not required to be prepared, and the prediction cost of the product demand is reduced.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of an application scenario of a product demand prediction method provided by an embodiment of the present application;

FIG. 2 is a schematic flow chart of a product demand prediction method according to an embodiment of the present application;

FIG. 3 is a schematic flow chart of the replenishment of a target product according to a demand provided in an embodiment of the present application;

FIG. 4 is a schematic flow chart of replenishment of the pre-bins provided in an embodiment of the application;

FIG. 5 is a schematic flow chart of selecting a target product according to a benefit amount from the candidate products according to the embodiment of the application

FIG. 6 is a schematic diagram showing the structure of an embodiment of a product demand prediction apparatus provided in an embodiment of the present application;

fig. 7 is a schematic structural diagram of an embodiment of an electronic device provided in an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to fall within the scope of the application.

In describing embodiments of the present application, it should be understood that the terms "first," "second," and "second" are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or number of features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more of the described features. In the description of the embodiments of the present application, the meaning of "plurality" is two or more, unless explicitly defined otherwise.

The following description is presented to enable any person skilled in the art to make and use the application. In the following description, details are set forth for purposes of explanation. It will be apparent to one of ordinary skill in the art that the present application may be practiced without these specific details. In other instances, well-known processes have not been described in detail in order to avoid unnecessarily obscuring the description of the embodiments of the application. Thus, the present application is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

The embodiment of the application provides a product demand prediction method, a product demand prediction device, electronic equipment and a readable storage medium. The product demand prediction device can be integrated in electronic equipment, and the electronic equipment can be a server or a terminal and other equipment.

The execution body of the product demand prediction method in the embodiment of the present application may be a product demand prediction device provided in the embodiment of the present application, or different types of electronic devices such as a server device, a physical host, or a User Equipment (UE) integrated with the product demand prediction device, where the product demand prediction device may be implemented in a hardware or software manner, and the UE may specifically be a terminal device such as a smart phone, a tablet computer, a notebook computer, a palm computer, a desktop computer, or a personal digital assistant (Personal Digital Assistant, PDA).

The electronic device may be operated in a single operation mode, or may also be operated in a device cluster mode.

Referring to fig. 1, fig. 1 is a schematic view of a product demand prediction system according to an embodiment of the present application. The product demand prediction system may include an electronic device 101, where a product demand prediction apparatus is integrated in the electronic device 101.

In addition, as shown in FIG. 1, the product demand prediction system may also include a memory 102 for storing data, such as text data.

It should be noted that, the schematic view of the scenario of the product demand prediction system shown in fig. 1 is only an example, and the product demand prediction system and the scenario described in the embodiments of the present application are for more clearly describing the technical solution of the embodiments of the present application, and do not constitute a limitation on the technical solution provided by the embodiments of the present application, and those skilled in the art can know that, with the evolution of the product demand prediction system and the appearance of a new service scenario, the technical solution provided by the embodiments of the present application is equally applicable to similar technical problems.

In the following, a product demand prediction method provided by an embodiment of the present application is described, where in the embodiment of the present application, an electronic device is used as an execution body, and in order to simplify and facilitate description, in a subsequent method embodiment, the execution body is omitted, and the product demand prediction method includes: acquiring historical sales data of a target product to be predicted; determining a product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time; selecting product sales data from the historical sales data according to the product sales characteristic value; and predicting the demand of the target product according to the product sales data.

Referring to fig. 2, fig. 2 is a schematic flow chart of a product demand prediction method according to an embodiment of the present application. It should be noted that although a logical order is depicted in the flowchart, in some cases the steps depicted or described may be performed in a different order than presented herein. The product demand prediction method specifically may include the following steps 201 to 204, where:

201. and acquiring historical sales data of the target product to be predicted.

The embodiment of the application does not limit the types of products, the products can be long-tail products such as automobile parts in the automobile parts industry, household appliance parts in the household appliance industry and the like, and the long-tail products are products with infrequent market demands and large market demand changes. Thus, the target product may be one of automobile engine, automobile tire, etc. parts, or one of refrigerator door, refrigerator refrigerating part, etc. parts in household appliances industry.

Sales data refers to data related to sales of products, for example, sales data may include time of sale of products, number of sales per day, and the like. The sales data may be obtained from data contained in the order, or a record of the product's shipment may be read from a physical management platform to obtain sales data for the product, for example. For example, the electronic device may calculate the conditions of all orders according to the time of placing orders on the orders corresponding to the target products and the sales volume corresponding to each order to obtain the historical sales data of the target products, or may query the daily sales volume of the target products from the physical management platform to obtain the historical sales data of the target products, and if not specifically described, the sales volume may be understood as the sales volume, and vice versa. For convenience of explanation, reference may be made to table 1, in which one case of historical sales data of a target product is shown in table 1:

TABLE 1

In other embodiments, to reduce the amount of computation, a portion of the sales data for the target product may be selected as historical sales data. For example, the electronic device may select a historical order between n months according to the time of placing the order corresponding to the target product, and count all the situations of the historical orders according to the time of placing the order on the historical order and the sales corresponding to each historical order, so as to obtain the historical sales data of the target product, where n may be set according to the needs.

202. And determining the characteristic value of the sales volume of the target product according to the sales time in the historical sales data and the sales volume of the product corresponding to each sales time.

The sales characteristic value is a value for describing characteristics of each data in sales data, and can be used to describe sales conditions of products corresponding to the sales data. For example, when the historical sales data includes the number of target products sold per day, the product sales feature value may be used to describe the time feature of the target product's date of sale, the change feature of the target product's daily sales, and so on.

For example, the product sales characteristic value may include the interval value of the selling date of the target product, the average value of the daily sales change of the target product, and the average value of the total sales of the target product, so as to accurately describe the sales situation of the target product. At this time, the step of determining the product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time may be obtained by:

(1) And determining the product selling time according to the selling time in the historical selling data and the product sales volume corresponding to each selling time.

(2) And determining the average value of the sales interval of the target product according to the sales time of each product.

The sales time refers to each time included in the historical sales data. Taking table 1 as an example for illustration, if the data in table 1 is historical sales data, the sales time refers to 1 month and 1 day to 1 month and 7 days. The time may be a date, or may be specifically an hour, a minute or a second, which is not limited in the embodiment of the present application, and for convenience of explanation, the time is hereinafter understood to be a date.

The product selling time refers to the time when the product is sold, and thus the product selling time in the historical sales data refers to the time when the target product is sold. In table 1, the sales of target products are generated on 1 month 2, 3, 4, 6, and 7 days, i.e., target products are sold on 1 month 2, 3, 4, 6, and 7 days, and thus, when only the data in table 1 is taken as the historical sales data, the sales time of the target products is 1 month 2, 3, 4, 6, and 7 days.

It will be appreciated that the sales time in the historical sales data includes the product sales time and the unproductive sales time, and the example of table 1 is taken as an example to continue the explanation, and in table 1, sales of target products are generated on 1 month 2 day, 3 day, 4 day, 6 day, and 7 day, and sales of target products are not generated on 1 month 1 day and 1 month 5 day, so that when only the data in table 1 is taken as the historical sales data, the product sales time refers to 1 month 2 day, 3 day, 4 day, 6 day, and 7 day, and the unproductive sales time refers to 1 month 1 day and 1 month 5 day.

The average value of the sales interval refers to the average value of the interval time between the sales times of the products, and can be understood as the average sales interval of the products. The following describes a method for calculating the average value of sales intervals, but this should not be taken as limiting the embodiments of the present application: taking the selling time of the product as an example, 1 month, 2 days, 3 days, 4 days, 6 days and 7 days, the latest time and the earliest time, namely 1 month, 7 days and 1 month, 2 days, can be selected from the selling times, then the time difference between the two times is obtained, and the number of the selling times of the product is divided to obtain the average value of the selling intervals. In this example, the average sales interval is 6/5 days, i.e., 1.2 days, and the average sales interval for the target product is 1.2 days. It can be appreciated that the sales interval average can be used to characterize market demand conditions for the target product, with larger sales interval averages indicating more frequent market demands for the target product.

The sales of the target product in the historical sales data refers to sales of the target product at each sales time. For example, when the sales time is calculated by day, the sales amount of the product in the historical sales data may refer to the number of target products sold per day, and taking the data in table 1 as the historical sales data, the sales amount of the product includes: 0,5,7,3,0,2,7. For ease of understanding, the sales of products in the historical sales data are understood hereinafter to be the number of target products sold per day, but are not to be construed as limiting embodiments of the application, if not specifically described. It can be understood that the time corresponding to the sales of the product is the product selling time.

(3) And determining a historical sales volume average value and a historical sales volume change average value of the target product according to the sales volume of each product.

The historical sales average refers to an average of sales of products, and if sales of products in the historical sales data are understood to be the number of target products sold per day, the historical sales average refers to the average number of target products sold per day. For example, when the sales of the product included 0,5,7,3,0,2,7 in table 1 and the respective dates are also consistent with table 1, the average of the historical sales is 3.4, i.e., the target product is between 1 month and 7 days, and the average sales per day is 3.4.

The historical sales volume change average value is the average value of change values among sales volumes of products, and the larger the historical sales volume change average value is, the larger the fluctuation of the number of target products sold every day is. For example, the average of historical sales changes may be derived from calculating standard deviations of sales of products. For example, when the product sales included 0,5,7,3,0,2,7 in table 1, the calculated average of the historical sales changes was 2.8.

(4) And determining the historical sales volume average value, the historical sales volume change average value and the sales interval average value as product sales volume characteristic values of the target product.

203. And selecting product sales data from the historical sales data according to the product sales characteristic value.

The product sales data is sales data selected from the historical sales data that accurately characterizes sales of the target product. The reason why the product sales data need to be selected instead of directly predicting by using the historical sales data is that when the target product is a long-tail product, the demand is not frequent and the demand amount is changed greatly, so that only data with smaller intervals between the corresponding time and the predicting time have reference significance to the prediction, conversely, data with larger intervals between the corresponding time and the predicting time may reduce the accuracy of the prediction due to not necessarily conforming to the market demand condition during the prediction, and it is seen that the data contained in the historical sales data has not necessarily conforming to the market demand condition during the prediction due to the large time span. On the other hand, if less data is used in the prediction, the accuracy of the prediction is also reduced, so that a reasonable amount of product sales data needs to be selected from the historical sales data according to the characteristic value of the product sales to improve the accuracy of the prediction.

For example, the target category of the target product may be first determined from a plurality of preset product categories according to the characteristic value of the sales volume of the product, and then the product sales data may be selected from the historical sales data according to a selection method corresponding to the target category. At this time, the step of selecting the product sales data from the historical sales data according to the product sales characteristic value may be performed by:

(a) And if the product sales characteristic value meets a preset first condition, counting the sales time of each product in the historical sales data to obtain the ex-warehouse frequency of the target product.

The first condition is a preset condition for judging whether the target product belongs to more frequent requirements. When the target product belongs to the long-tail product and the demand is more frequent, the prediction inaccuracy greatly influences the supply condition of a manufacturer to a customer, so that the demand on the prediction accuracy is higher compared with the long-tail product with less demand. At this time, proper product sales data is selected from the historical sales data according to the characteristic value of the sales volume of the product, so as to improve the prediction accuracy.

The preset first condition can be set according to scene requirements, and the requirement frequency degree of the target product can be reasonably judged. The following provides a first condition setting manner, but is not to be construed as limiting the embodiment of the present application, and it is assumed that the product sales characteristic value of the target product includes a historical sales average, a historical sales change average and a sales interval average:

(a1) Firstly, calculating according to a historical sales volume average value and a historical sales volume change average value in a product sales volume characteristic value to obtain a variation coefficient (Coefficient of Variation) of historical sales volume data, wherein the variation coefficient is a ratio between the historical sales volume change average value and the historical sales volume average value, reflects the relative discrete degree between the product sales volumes, and can be referred to above for the description of the product sales volume without being repeated.

(a2) And respectively comparing the variation coefficient, the historical sales volume average value and the sales interval average value with a preset first threshold value, a preset second threshold value and a preset third threshold value, and if the variation coefficient is smaller than or equal to the first threshold value, the historical sales volume average value is larger than or equal to the second threshold value and the sales interval average value is smaller than or equal to the third threshold value, indicating that the characteristic value of the sales volume of the product meets a first condition, and indicating that the sales volume of the product of the target product is large enough and the demand is frequent enough. Wherein the first threshold, the second threshold, and the third threshold may be set according to scene requirements, for example, the first threshold may be set to 10, the second threshold may be set to 2, and the third threshold may be set to 60.

The ex-warehouse frequency refers to the number of times the target product is ex-warehouse, i.e. the number of times the target product is sold. For example, the sales time in the historical sales data may be divided into a plurality of time periods according to a preset time interval, and then the number of time periods including the product sales time in each time period is counted to obtain the ex-warehouse frequency, where the product sales time refers to the time of product sales, and reference may be made to the above description. For example, the time interval may be set to 7 days, i.e. counting the number of weeks containing the time of sale of the product. In the following explanation, if the time interval is set to 7 days, the sales time in the historical sales data is 1 month 1 day to 1 month 21 days, and the product sales time includes 1 month 1 day, 1 month 2 day, and 1 month 13 day, the 1 month 1 day to 1 month 21 day is first divided into 3 weeks according to the time interval: 1 month 1 day to 1 month 7 day, 1 month 8 day to 1 month 14 day, 1 month 15 day to 1 month 21 day, respectively referred to as the first week, the second week and the third week, and then the weeks including 1 month 1 day, 1 month 2 day and 1 month 13 day were judged to be the first week and the second week, so that the frequency of delivery was 2. Or, the time interval may be set to 1 day, and the statistics of the frequency of sales of the product is the statistics of the number of sales times of the product, which is not illustrated in detail.

(b) If the ex-warehouse frequency is smaller than a preset frequency threshold, dividing the sales time in the historical sales data into preset time periods, counting the total product sales in each preset time period, and taking the ex-warehouse time period and the corresponding product sales which are not zero as the product sales data.

The preset time period is a time period obtained by dividing the sales time in the historical sales data according to a preset time interval, and the dividing method in the step (a) may be referred to, which is not described in detail, and hereinafter, for convenience of explanation, the default time interval is 7 days.

The ex-warehouse time period refers to a time period in which the total value of sales of corresponding products is not 0 in a preset time period, that is, if the sales of a target product in at least one day is not 0 in 7 days corresponding to one preset time period, the preset time period can be taken as an ex-warehouse time period.

In the following explanation, if the time interval is set to 7 days, the historical sales data contains product sales data of 1 month, 1 day, and 1 month, and 21 days, and the product sales time includes 1 month, 1 day, 1 month, 2 days, and 1 month, 13 days, the 1 month, 1 day, and 1 month, 21 days are first divided into 3 weeks according to the time interval: 1 month 1 day to 1 month 7 day, 1 month 8 day to 1 month 14 day, 1 month 15 day to 1 month 21 day, respectively referred to as a first week, a second week and a third week, the first week, the second week and the third week are respectively a preset time period, and the first week and the second week are respectively a time period of leaving a warehouse because the sales of the products on 1 month 1 day, 1 month 2 day and 1 month 13 day are not 0, and the obtained sales data of the products are respectively historical sales data of corresponding time between 1 month 1 day and 1 month 14 day.

If the ex-warehouse frequency is smaller than a preset frequency threshold, the historical sales data contains sales data of the target product with the market demand in the valley period, wherein the frequency threshold can be set according to the scene demand, for example, the frequency threshold can be set to be 9. When the target product belongs to a long-tail product and the demand is frequent (the preset first condition is met), if sales data obtained when the market demand of the target product is in the valley period is not removed during prediction, the demand of the predicted target product may be obviously smaller than the actual demand in the future, so that the purchasing demand of a customer cannot be met. And after the historical sales data corresponding to the ex-warehouse time period is used as the product sales data, the cycle time when the market demand of the target product is in the valley period can be removed, the historical sales data corresponding to the cycle time when the market demand of the target product is in the normal state is used as a prediction basis, the prediction accuracy is improved, and the situation that the purchasing demand of a customer cannot be met is avoided.

(c) If the ex-warehouse frequency is greater than or equal to a preset frequency threshold, comparing each sales time in the historical sales data with a preset time condition to obtain a target sales time meeting the preset time condition, and taking the target sales time and the corresponding product sales volume as product sales data.

The preset time condition may be a condition for judging whether or not the sales time in the historical sales data is sufficiently close to the time at the time of prediction. For example, a time difference between the sales time in the historical sales data and the time at the time of prediction may be less than or equal to a preset time threshold, which may be set according to scene needs, as a requirement to satisfy the preset time condition.

If the ex-warehouse frequency is greater than or equal to the preset frequency threshold, it is indicated that the historical sales data does not contain sales data of the target product with the market demand in the valley period, so that the sales data can be not removed, but the sales data is taken as the sales data of the product with the corresponding sales time being close enough to the time of the forecast, namely the latest historical sales data is taken as the sales data of the product, so that the forecast of the demand is matched with the market demand condition of the forecast, and the accuracy of the forecast is improved. In the following, an example will be described as an example, assuming that the time at the time of prediction is 2022, 1 month and 10 days, and only the data in table 1 is taken as the history sales data, and the time difference between the time and the time at the time of prediction is less than or equal to 5 days as required to satisfy the preset time condition, then only the sales time of 1 month, 5 days, 6 days, 7 days, and the corresponding sales amount of the product are taken as the product sales data.

On the other hand, if the characteristic value of the sales volume of the product does not meet the preset first condition, the market demand of the target product is indicated to be low-frequency, and at the moment, the historical sales data can be directly used as the sales data of the product, and the situation that the purchasing demand of a customer cannot be met is avoided, so that selection is not needed, and the calculated amount can be reduced.

204. And predicting the demand of the target product according to the product sales data.

In the embodiment of the application, various methods can be adopted to predict the demand of the target product.

In some embodiments, machine Learning (Machine Learning) methods may be used to predict the demand for the target product. For example, the sample sales data may be used as training data to train an initial logistic regression (logistic regression) model to obtain a trained logistic regression model, and then the product sales data is input into the trained logistic regression model to predict the demand of the target product.

Machine learning is a simulated human learning approach that uses a computer as a tool and is directed to real time.

The logistic regression model is a generalized linear regression analysis model and is commonly used in the fields of data mining, automatic disease diagnosis, economic prediction and the like.

However, the method of applying machine learning is extremely dependent on the representativeness of the training data, and if the training data cannot represent the market demand situation of the target product, the predicted demand of the target product may be inaccurate.

In other embodiments, therefore, the sales average and sales change average of the target product may be calculated again according to the product sales data, and the demand of the target product may be predicted according to the obtained sales average and sales change average. The sales average value and the sales change average value are obtained based on the product sales data of the target product, so that the market demand condition of the target product can be represented, and the demand of the target product obtained through prediction is more accurate. At this time, the step of "predicting the demand amount of the target product from the product sales data" may be obtained by:

(A) And determining a target sales volume average value and a target sales volume change average value of the target product according to the sales volume of each product in the product sales data.

The method for calculating the target sales average and the target sales variation average can refer to the above method. In the prediction process, the target sales volume average value is based on the past sales data of the target product, can be used for representing the future average demand volume of the target product, and the target sales volume change average value is also based on the past sales data of the target product, can be used for representing the future demand volume fluctuation condition of the target product, and the demand volume obtained through the target sales volume average value and the target sales volume change average value prediction not only accords with the actual market demand condition of the target product, but also considers the possible fluctuation condition of the demand volume, so that compared with a machine learning method, the obtained demand volume is accurate, and the situation that the purchasing demand of customers cannot be satisfied is avoided.

It should be noted that, if the product sales data is selected by the method from step (a) to step (c), when the frequency of delivery is smaller than the preset frequency threshold, the target sales average value and the target sales variation average value may be calculated in different manners according to the size of the frequency of delivery, so as to improve the accuracy of prediction. For example, a preset high-frequency threshold value and a preset low-frequency threshold value may be further set, and different processes are performed according to the magnitude of the frequency of delivery, which are exemplarily described below:

(A1) If the frequency of ex-warehouse is smaller than the preset frequency threshold value and larger than the high-frequency threshold value, the target sales average value and the target sales change average value can be calculated according to the method.

(A2) If the ex-warehouse frequency is smaller than or equal to the high-frequency threshold value and larger than the low-frequency threshold value, the target sales volume average value can be calculated according to the method, and because the target sales volume change average value is used for representing the future requirement volume fluctuation condition of the target product, the situation that a large amount of target products with the ex-warehouse frequency smaller than the high-frequency threshold value suddenly have a large number of requirements is less common, in order to reduce the quantity of the stagnant inventory as much as possible, the target sales volume change average value can be set as a lower change average value threshold value, the change average value threshold value can be set according to scene requirements, for example, the change average value threshold value can be set as 0, namely, the quantity of the stagnant inventory is reduced as much as possible without considering the future requirement volume fluctuation condition of the target products.

Wherein, the dead stock refers to stock products which are stored for a long time and cannot be sold.

(A3) If the frequency of ex-warehouse is less than or equal to the low frequency threshold, in order to further reduce the number of dead warehouses, the initial sales average value may be calculated first according to the method in step (A1), and then the initial sales average value is multiplied by a preset coefficient threshold smaller than 1 to obtain the target sales average value, where the preset coefficient threshold may be set according to the scene requirement, for example, the preset coefficient threshold may be set to 0.5. For the target sales change mean value, the same processing method as that in the step (A2) may be adopted, and detailed description is omitted.

The high-frequency threshold and the low-frequency threshold in the steps (A1) to (A3) may be set according to the scene requirement, for example, when the frequency threshold is 9, the high-frequency threshold may be set to 5, and the low-frequency threshold may be set to 1.

(B) And determining a target service level corresponding to the target product according to the product sales characteristic value.

The service level is an inventory service level (Inventory service level) indicating that the target product is at x% backorder when the service level of the target product is at x%.

For example, it may be first determined whether the product sales feature value meets a preset first condition, and the target service level is selected from a plurality of preset service levels according to the determination result, where the first condition in step (B) may be understood as the first condition in step (a) above, that is, the first condition is a preset condition for determining whether the target product belongs to a more frequent demand, which is not specifically illustrated. In the following, an exemplary explanation is given by way of example, 2 preset service levels may be set, and when the product sales characteristic value satisfies the first condition, the first preset service level is taken as the target service level. And when the product sales characteristic value does not meet the first condition, setting the second preset service level as a target service level. It will be appreciated that when the market demand of the target product is more frequent, the corresponding service level should be higher to avoid the occurrence of an out-of-stock condition, and thus the first preset service level should be greater than the second preset service level, for example, the first service level may be set to 95% and the second preset service level may be set to 80%.

(C) And determining the safety stock value of the target product according to the target service level and the target sales volume change mean value.

Illustratively, the safety stock value can be calculated by equation (1):

q=kσ type son (1)

Wherein Q is a safety stock value, sigma is a target sales volume change mean value, k is a Z value obtained by taking the target sales volume change mean value as a standard deviation and taking a target service level as an upper alpha quantile in target normal distribution constructed by taking the target sales volume mean value as a mathematical expectation (mathematic expectation).

In step (C), the safety stock value refers to a future maximum demand fluctuation value of the target product.

(D) And predicting the demand of the target product according to the safety stock value and the target sales average value.

Since the safety stock value refers to the future maximum demand fluctuation value of the target product, and the target sales average value is based on the past sales data of the target product and can be used for representing the future average demand of the target product, the demand of the target product can be predicted according to the safety stock value and the target sales average value by the formula (2):

demand=μ+q type son (2)

Wherein demand is the predicted demand of the target product, mu is the target sales average value, and Q is the safety stock value.

In summary, the product demand prediction method provided by the embodiment of the application includes: acquiring historical sales data of a target product to be predicted; determining a product sales characteristic value of the target product according to the sales time in the historical sales data and the product sales corresponding to each sales time; selecting product sales data from the historical sales data according to the product sales characteristic value; and predicting the demand of the target product according to the product sales data. On one hand, the product demand prediction method provided by the embodiment of the application selects the product sales data based on the product sales characteristic value of the target product, and because the product sales characteristic value contains the change relation of the product sales along with the sales time, the product demand prediction method can be used for judging whether the historical sales data can correctly reflect the market demand condition of the current target product, and further selecting the effective product sales data from the historical sales data when the historical sales data contains redundancy data which cannot be correctly reflected, compared with the method for predicting by directly adopting the historical sales data, the obtained demand of the target product is more accurate, and the quantity of dead stock can be effectively reduced according to the product characteristics of the automobile industry. On the other hand, the product demand prediction method provided by the embodiment of the application does not depend on the training of sample data, so that compared with a method for predicting by adopting machine learning, the prediction error generated by the fact that the sample data is not representative can be avoided, and the predicted data base is the historical sales data of the target product, so that a large amount of sample data is not required to be prepared, and the prediction cost of the product demand is reduced.

In some embodiments, after the demand of the target product is predicted, the target product may be restocked according to the obtained demand, so as to avoid that the inventory of the target product in the warehouse is insufficient to cope with the purchase demand of the customer. In order to minimize the quantity of dead stock, replenishment can be performed only when the stock value in the warehouse is lower than the safety stock value when the market demand of the target product is determined to be infrequent according to the product sales characteristic value. Referring to fig. 3, at this time, after the step of "predicting the required amount of the target product from the product sales data", the method further includes:

301. and if the product sales characteristic value does not meet a preset first condition, acquiring a target inventory value of the target product.

For ease of understanding, when the product sales feature value does not satisfy the first condition, it is referred to as the product sales feature value satisfying the second condition.

Taking the requirement of the first condition as the requirement of the steps (a 1) to (a 2) as an example, the requirement of the first condition is as follows: if the coefficient of variation is less than or equal to a first threshold, the average of the historical sales volume is greater than or equal to a second threshold, and the average of the sales intervals is less than or equal to a third threshold, (1) the coefficient of variation is greater than the first threshold; (2) The coefficient of variation is less than or equal to a first threshold and the historical sales mean is less than a second threshold; (3) The coefficient of variation is less than or equal to a first threshold, the average of the historical sales volume is greater than or equal to a second threshold, and the average of the sales interval is greater than a third threshold; all of the above three cases are cases where the second condition is satisfied. And when the product sales feature meets a preset second condition, indicating that the market demand of the target product is not frequent.

The target inventory value of the target product refers to the amount the target product has currently stored in the warehouse. For example, if the market demand for the target product is infrequent, the target of restocking may be a central warehouse, and thus the target inventory value may refer to the amount the target product has currently stored in the central warehouse. The central warehouse is a warehouse which is communicated with a product delivery base and communicated with the market end, can be arranged in large cities such as first-line cities, and can be used for retrieving a corresponding number of target products from the central warehouse to deliver the products to the cities when the cities without the central warehouse generate the purchasing demands of the target products, so that the construction cost of the warehouse can be reduced by storing the target products in the central warehouse.

For example, the electronic device may query a warehouse management system of the central warehouse to obtain a target inventory value for the target product.

302. And if the target stock value is smaller than or equal to the safety stock value of the target product, replenishing according to the demand, wherein the safety stock value is obtained according to the preset service level corresponding to the product sales characteristic value and the product sales data.

If the target stock value is less than or equal to the safe stock value of the target product, the target product stored in the warehouse is insufficient to meet the purchase requirement of the customer, and if the stock is not restocked, the stock shortage condition may occur, so that the stock is restocked according to the requirement. The method for obtaining the safety stock value may refer to the above, and will not be described herein.

For example, a difference between the demand and the target inventory value may be calculated and restocking performed on the warehouse according to the difference.

In some embodiments, a pre-warehouse may be provided between the central warehouse and the marketplace to quickly respond to customer purchase needs. Referring to fig. 4, the replenishment method at this time may be:

401. and counting the historical sales volume ratio of the target product in each preset warehouse.

The pre-set warehouse in step 401 may refer to a pre-set warehouse set between the central warehouse and the market.

The historical sales ratio of the target products in one preset warehouse refers to the ratio of the total sales value of the corresponding target products in the preset warehouse to the total sales value of the target products in all preset warehouses. For example, the total sales value of the target product in all preset warehouses is 100 pieces, wherein when the total sales value of the target product in the preset warehouse 1 accounts for 50 pieces, the historical sales ratio of the target product in the preset warehouse 1 accounts for 50%.

In some embodiments, the electronic device may read a warehouse management system of each preset warehouse to obtain a historical sales rate of the target product in each preset warehouse.

402. Dividing the demand according to the historical sales volume occupation ratio of each preset warehouse to obtain the replenishment volume of each preset warehouse.

For ease of understanding, step 402 is exemplarily described below with reference to table 2, where the historical sales ratios for each of the preset warehouses are shown in table 2:

table 2 assuming that the required amount is 100, the restocking amount of each preset warehouse can be referred to table 3:

preset warehouse	Preset warehouse 1	Preset warehouse 2	Preset warehouse 3	Preset warehouse 4	Preset warehouse 5
						Quantity of goods supplement	20	30	10	20	20

TABLE 3 Table 3

403. And replenishing the goods in each preset warehouse according to the replenishing quantity of each preset warehouse.

It should be noted that, steps 401-403 may be performed only when the market demand of the target product is more frequent, that is, the product sales feature value satisfies the first condition, otherwise steps 301-302 may be performed. The description of the first condition may refer to the above, and detailed description is omitted.

In some embodiments, demand forecasting may be performed only on profitable products to increase utilization of computing resources. Referring to fig. 5, at this time, step "acquire historical sales data of a target product to be predicted" includes:

501. Sales data for a plurality of alternative products is obtained.

Wherein the alternative products may be all possible product categories. For example, in the case of demand prediction for parts in the automotive industry, an alternative product refers to a part in the automotive industry, and may include tires, doors, engines, and the like.

The description of sales data may be referred to above, and detailed description thereof will be omitted.

502. And determining the income amount of each alternative product according to the preset product cost and the preset sales amount corresponding to each alternative product.

The preset product cost corresponding to an alternative product refers to the cost required for each of the alternative products in the production of such alternative products. Which may include labor costs, material costs, and the like.

The preset sales amount of an alternative product refers to the price of the alternative product when sold to a customer.

The amount of benefit for an alternative product refers to the amount of benefit per sale of one such alternative product.

The profit amount of an alternative product can be calculated according to the preset sales amount and the preset product cost of the alternative product.

503. And screening each candidate product according to the profit amount of each candidate product to obtain a target product with the corresponding profit amount larger than a preset value threshold, and taking the sales data of the target product as historical sales data.

The preset monetary threshold is a preset value for evaluating the amount of the return, and when the corresponding amount of the return is greater than the preset monetary threshold, the candidate product can be considered to have a high profitability, and whether the candidate product needs to be restocked in time is determined so as to avoid the occurrence of a stock shortage condition.

In addition to selecting the target product according to the profit amount by the method of step 501-step 503, the candidate product may be selected according to the number of the product selling time in the sales data, and the number of the corresponding product selling time is greater than the preset number threshold value to be used as the target product. The preset number threshold can be set according to scene requirements.

In some embodiments, the target product may be classified according to the product sales feature value of the target product to obtain a target type of the target product, and then product sales data is selected from the historical sales data according to the target type, so as to predict the demand of the target product, and the nature and steps of the product sales data are consistent with those of the target product, so that the product sales data are not additionally described, and only an example is given below to facilitate understanding:

(i) The target product is selected from the candidate products by a screening method of the profit amount and the product selling time, and the specific description of the screening method can be referred to above, and details are not repeated.

(ii) Historical sales data of the target product is obtained from channels such as orders, logistics management platforms and the like.

(iii) And (3) obtaining the product sales characteristic value of the target product by the methods of the steps (1) to (4).

(iv) Calculating a variation coefficient according to the historical sales volume average value and the historical sales volume change average value, and comparing the variation coefficient, the historical sales volume average value and the sales interval average value with a first threshold value, a second threshold value and a third threshold value respectively, wherein the meaning of the first threshold value, the second threshold value and the third threshold value is the same as that of the step (a 2).

On the one hand, if the variation coefficient is smaller than or equal to the first threshold value, the historical sales volume average value is larger than or equal to the second threshold value, and the sales interval average value is smaller than or equal to the third threshold value, removing zero in the sales volume of the product, calculating to obtain a non-zero sales volume average value and a non-zero sales volume change average value according to the remaining non-zero sales volume, calculating to obtain a non-zero variation coefficient according to the non-zero sales volume average value and the non-zero sales volume change average value, setting the target type to be high-frequency stable if the non-zero variation coefficient of the target product is smaller than or equal to a preset variation coefficient threshold value, and setting the target type to be high-frequency unstable if the non-zero variation coefficient of the target product is larger than the preset variation coefficient threshold value.

On the other hand, if:

(1) The variation coefficient is larger than a first threshold value, and the target type is set to be extremely unstable;

(2) The variation coefficient is smaller than or equal to a first threshold value, and the average value of the historical sales is smaller than a second threshold value, the target type is set to be extremely small;

(3) The coefficient of variation is less than or equal to a first threshold, the average value of the historical sales volume is greater than or equal to a second threshold, and the average value of the sales interval is greater than a third threshold, the non-zero coefficient of variation is also calculated, if the non-zero coefficient of variation of the target product is less than or equal to a preset coefficient of variation threshold, the target type is set to be low-frequency stable, and if the non-zero coefficient of variation of the target product is greater than the preset coefficient of variation threshold, the target type is set to be low-frequency unstable.

(v) If the target type is high frequency stable or high frequency unstable, selecting product sales data from the historical sales data by the method of step (a) -step (c). If the target type is one of extremely unstable, extremely small, low frequency stable, and low frequency unstable, the historical sales data is taken as product sales data.

(vi) The target sales volume average value and the target sales volume change average value are calculated according to the product sales data, and a specific calculation method is not repeated.

(vii) And (3) predicting the demand of the target product by the method from the step (A) to the step (D).

(viii) If the target type is high frequency stable or high frequency unstable, the pre-bins are restocked by the method of steps 401-403. If the target type is one of extremely unstable, extremely small, low frequency stable and low frequency unstable, the central warehouse is restocked only when the target inventory value is less than or equal to the safety inventory value corresponding to the target product.

In order to better implement the product demand prediction method in the embodiment of the present application, on the basis of the product demand prediction method, the embodiment of the present application further provides a product demand prediction device, as shown in fig. 6, which is a schematic structural diagram of an embodiment of the product demand prediction device in the embodiment of the present application, where the product demand prediction device 600 includes:

an obtaining unit 601, configured to obtain historical sales data of a target product to be predicted;

a determining unit 602, configured to determine a product sales characteristic value of the target product according to a sales time in the historical sales data and a product sales volume corresponding to each sales time;

a selecting unit 603, configured to select product sales data from the historical sales data according to the product sales feature value;

And a prediction unit 604, configured to predict a demand of the target product according to the product sales data.

In a possible implementation of the present application, the determining unit 602 is further configured to:

In a possible implementation of the application, the selection unit 603 is further configured to:

In a possible implementation of the present application, the prediction unit 604 is further configured to:

In a possible implementation of the present application, the obtaining unit 601 is further configured to:

acquiring sales data of a plurality of alternative products;

In the implementation, each unit may be implemented as an independent entity, or may be implemented as the same entity or several entities in any combination, and the implementation of each unit may be referred to the foregoing method embodiment, which is not described herein again.

Since the product demand prediction device can execute steps in the product demand prediction method in any embodiment, the product demand prediction method in any embodiment of the present application can achieve the beneficial effects, which can be achieved by the product demand prediction method in any embodiment of the present application, and detailed descriptions are omitted herein.

In addition, in order to better implement the product demand prediction method in the embodiment of the present application, on the basis of the product demand prediction method, the embodiment of the present application further provides an electronic device, referring to fig. 7, fig. 7 shows a schematic structural diagram of the electronic device in the embodiment of the present application, and specifically, the electronic device provided in the embodiment of the present application includes a processor 701, where the processor 701 is configured to implement each step of the product demand prediction method in any embodiment when executing a computer program stored in a memory 702; alternatively, the processor 701 is configured to implement the functions of the respective modules in the corresponding embodiment of fig. 6 when executing the computer program stored in the memory 702.

By way of example, a computer program may be partitioned into one or more modules/units that are stored in the memory 702 and executed by the processor 701 to accomplish an embodiment of the application. One or more of the modules/units may be a series of computer program instruction segments capable of performing particular functions to describe the execution of the computer program in a computer device.

Electronic devices may include, but are not limited to, processor 701, memory 702. It will be appreciated by those skilled in the art that the illustrations are merely examples of electronic devices and are not limiting of electronic devices, and may include more or fewer components than illustrated, or may combine certain components, or different components.

The processor 701 may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like, which is a control center for an electronic device, with various interfaces and lines connecting various parts of the overall electronic device.

The memory 702 may be used to store computer programs and/or modules, and the processor 701 implements the various functions of the computer device by running or executing the computer programs and/or modules stored in the memory 702, and invoking data stored in the memory 702. The memory 702 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like; the storage data area may store data (such as audio data, video data, etc.) created according to the use of the electronic device, and the like. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the product demand prediction apparatus, the electronic device and the corresponding units described above may refer to the description of the product demand prediction method in any embodiment, and will not be repeated herein.

Those of ordinary skill in the art will appreciate that all or a portion of the steps of the various methods of the above embodiments may be performed by instructions or by controlling associated hardware, which may be stored on a readable storage medium and loaded and executed by a processor.

Therefore, an embodiment of the present application provides a readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, performs steps in a product demand prediction method in any embodiment of the present application, and specific operations may refer to descriptions of the product demand prediction method in any embodiment, which are not described herein.

Wherein the readable storage medium may include: read Only Memory (ROM), random access Memory (RAM, random Access Memory), magnetic or optical disk, and the like.

The steps in the product demand prediction method in any embodiment of the present application may be executed by the instructions stored in the readable storage medium, so that the beneficial effects that the product demand prediction method in any embodiment of the present application may achieve may be achieved, which is described in detail in the foregoing, and will not be repeated herein.

The above description of the product demand prediction method, the device, the storage medium and the electronic equipment provided by the embodiment of the present application applies specific examples to describe the principles and the implementation of the present application, and the description of the above embodiments is only used to help understand the method and the core idea of the present application; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in light of the ideas of the present application, the present description should not be construed as limiting the present application.

Claims

1. A method of product demand prediction, comprising:

acquiring historical sales data of a target product to be predicted;

2. The product demand prediction method according to claim 1, wherein the determining the product sales characteristic value of the target product according to the sales time and the product sales amount corresponding to each sales time in the historical sales data includes:

Determining the selling time of the product according to the selling time in the historical selling data and the product sales corresponding to each selling time;

3. The product demand prediction method according to claim 1, wherein selecting product sales data from the historical sales data according to the product sales feature value comprises:

4. The product demand prediction method according to claim 1, wherein predicting the demand amount of the target product based on the product sales data comprises:

5. The product demand prediction method according to claim 1, wherein after predicting the demand amount of the target product based on the product sales data, the method further comprises:

6. The product demand prediction method according to claim 1, wherein after predicting the demand amount of the target product based on the product sales data, the method further comprises:

7. The method of claim 1-6, wherein the obtaining historical sales data of the target product to be predicted comprises:

acquiring sales data of a plurality of alternative products;

8. A product demand prediction apparatus, comprising:

9. An electronic device comprising a processor, a memory and a computer program stored in the memory and executable on the processor, the processor implementing the steps in the product demand prediction method according to any one of claims 1 to 7 when the computer program is executed by the processor.

10. A readable storage medium, characterized in that it has stored thereon a computer program which, when executed by a processor, implements the steps of the product demand prediction method according to any one of claims 1 to 7.