CN111275505A

CN111275505A - Data processing method and device, electronic equipment and storage medium

Info

Publication number: CN111275505A
Application number: CN202010380027.3A
Authority: CN
Inventors: 周鹏程; 崔燕达
Original assignee: Ainnovation Nanjing Technology Co ltd
Current assignee: Ainnovation Nanjing Technology Co ltd
Priority date: 2020-05-08
Filing date: 2020-05-08
Publication date: 2020-06-12

Abstract

The application relates to a data processing method, a data processing device, electronic equipment and a storage medium, and belongs to the technical field of computers. The method comprises the following steps: obtaining a decision target, a decision variable and a limiting condition which influence the profit maximization; the decision target is the return on investment or the total volume of the transaction; the decision variables comprise a product list participating in promotion and respective prices of N products, and the number of the products in the product list is not more than N; obtaining a sales prediction result corresponding to the price according to the price of each product in the N products; and solving an output result for maximizing the profit according to the sales prediction result, the decision factor and the optimization algorithm corresponding to the price of each of the N products. According to the method, a sales promotion product list capable of achieving the maximum profit and the price of each product are solved according to the sales prediction results of each product at different prices and the decision factors influencing the profit maximization, and the product selection and pricing aiming at the fine promotion of the online products or the offline products are realized.

Description

Data processing method and device, electronic equipment and storage medium

Technical Field

The present application belongs to the field of computer technologies, and in particular, to a data processing method and apparatus, an electronic device, and a storage medium.

Background

Promotion is one of the basic strategies for marketing, and good promotion can bring huge benefits. With the development of artificial intelligence, more and more artificial intelligence technologies are applied to promotion pricing optimization in various fields. Most of the existing offline or online sales promotion schemes mainly adopt several ways:

(1) the static information or transaction information of the user or the commodity is divided, and different promotion rules are formulated for each subdivided commodity. The disadvantages of this approach are mainly: the rules need to be customized according to different scenes of different commodities, and the commodities are difficult to migrate to different time and different scenes, so that the applicability is poor, and the efficiency is low.

(2) According to different market research results, different prior knowledge models are formulated, and then the parameters of the formulated models are fitted by using data. The disadvantages of this approach are mainly: 1) excessive market research and professional knowledge are required, so that much time is required for early preparation, and the efficiency is influenced; 2) any prior knowledge model has the limiting conditions, most of the practical situations can not meet all the limiting conditions, so the prior knowledge model can not completely represent the data situation, the actual landing effect can be influenced, and the prediction precision is poor.

Disclosure of Invention

In view of this, an object of the present application is to provide a data processing method, an apparatus, an electronic device, and a storage medium, so as to solve the problems of low efficiency and low prediction accuracy of the existing promotion scheme.

The embodiment of the application is realized as follows:

in a first aspect, an embodiment of the present application provides a data processing method, including: when an optimal promotion strategy that N products participating in promotion pricing are predicted to maximize benefits is needed, obtaining a decision factor influencing the benefits maximization, wherein the decision factor comprises: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target; obtaining a sales forecasting result corresponding to the price according to the price of each product in the N products; solving an output result for maximizing the profit according to the sales forecasting result, the decision factor and the optimization algorithm corresponding to the price of each of the N products, wherein the output result comprises: a product list and prices of each product in the product list. In the embodiment of the application, during sales volume prediction, prices are taken into consideration, different sales volumes can be predicted according to different prices, then a sales promotion product list capable of achieving the maximum profit and prices of products are solved according to the sales volume prediction results of the products at different prices and decision factors influencing profit maximization, then product type selection and pricing aiming at refined sales promotion of online products or offline products are achieved, and the problems of low efficiency and low prediction precision caused by poor applicability of the existing sales promotion scheme are solved.

With reference to a possible implementation manner of the embodiment of the first aspect, when the decision target is a return on investment ROI, the decision variables further include: the advertisement input of each of the N products, and accordingly, the limitation further includes: the advertising input fluctuation range of each product in the product list, and the output result further includes: advertising investment for each product in the product list; correspondingly, according to the price of each product in the N products, obtaining a sales prediction result corresponding to the price, including: obtaining a sales prediction result corresponding to the price and the advertisement investment according to the price and the advertisement investment of each product in the N products; correspondingly, solving an output result for maximizing the profit according to the sales prediction result corresponding to the price of each of the N products, the decision factor and the optimization algorithm, and comprises the following steps: and solving an output result for maximizing the profit according to the price of each of the N products, the sales prediction result corresponding to the advertisement investment, the decision factor and the optimization algorithm. In the embodiment of the application, when the decision target is the return on investment rate ROI, the decision variables of the ROI also can comprise respective advertisement investment of N products, so that the system can solve a promotion product list which can maximize the ROI and the price and the advertisement investment of each product without re-formulating the regulation, and the applicability of the scheme is expanded.

With reference to a possible implementation manner of the embodiment of the first aspect, the obtaining a decision factor that affects profit maximization includes: and responding to the configuration operation of the user to obtain a decision factor influencing the profit maximization. In the embodiment of the application, when the decision factor influencing the maximum income is obtained, the configuration operation of the user is responded in real time, and the latest decision factor is obtained, so that the solved output result is more accurate and closer to the requirement of the user.

With reference to a possible implementation manner of the embodiment of the first aspect, before obtaining, according to the price of each of the N products, a sales prediction result corresponding to the price, the method further includes: acquiring original data of the N products related to sales volume; aiming at each product, determining the category of the product and acquiring input characteristics according to the original data of the product, wherein the input characteristics comprise sales data, price and advertisement input information; and predicting the sales volume of the product by using the sales volume prediction model corresponding to the category and the input characteristics. In the embodiment of the application, the sales prediction results of the N products under different prices are obtained in advance, so that when the N products participating in promotion pricing need to be predicted to achieve the optimal promotion strategy of maximizing the income, the results are directly called to improve the running speed, and meanwhile, when the sales results are predicted, different product categories adopt different sales prediction models to improve the accuracy of efficiency prediction as much as possible.

With reference to a possible implementation manner of the embodiment of the first aspect, determining, according to the raw data of the product, a category to which the product belongs includes: and clustering each product based on the respective original data of the N products to obtain the category of each product. In the embodiment of the application, each product is clustered based on respective original data of N products, so that the category of each product can be quickly determined, and the operation efficiency is improved.

In combination with a possible implementation manner of the embodiment of the first aspect, obtaining raw data related to sales volume of each of the N products participating in the pricing for promotion includes: obtaining product data, transaction data, auction data, and other data relating to sales for each of the N products participating in promotional pricing, wherein the other data comprises, when the N products are all online products: basic information of a product purchaser, shopping cart data of a shopping cart into which the product is added, and behavior data including browsing and returning behaviors of the product purchaser for the product; when the N products are offline products, the other data includes: store data of the store to which the product belongs. In the embodiment of the application, when the original data of the product is obtained, the obtained original data are different aiming at two different scenes of the online product and the offline product, and the targeted data are obtained through different scenes, so that the targeted data can represent a real scene more, and the predicted sales amount is closer to reality and more accurate.

With reference to a possible implementation manner of the embodiment of the first aspect, before predicting the sales volume of the product by using the sales volume prediction model corresponding to the category and the input feature, the method further includes: acquiring original data of M products related to sales volume respectively, wherein M is an integer and is larger than N; clustering the M products based on respective original data of the M products to obtain a plurality of clustering categories; and respectively constructing a sales prediction model for each cluster category. In the embodiment of the application, products are clustered based on a large amount of original data of the products to obtain different clustering categories, and then sales prediction models are respectively constructed aiming at the different clustering categories, so that the established sales prediction models are more pertinent and representative, and the predicted sales are more accurate.

In a second aspect, an embodiment of the present application further provides a data processing apparatus, including: the system comprises an acquisition module, a calling module and a solving module; an obtaining module, configured to obtain a decision factor influencing revenue maximization when an optimal promotion strategy that maximizes revenue by predicting N products participating in promotion pricing is needed, where the decision factor includes: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target; the calling module is used for acquiring a sales forecasting result corresponding to the price according to the price of each product in the N products; a solving module, configured to solve an output result that maximizes a benefit according to the sales prediction result, the decision factor, and the optimization algorithm corresponding to the price of each of the N products, where the output result includes: a product list and prices of each product in the product list.

In a third aspect, an embodiment of the present application further provides an electronic device, including: a memory and a processor, the processor coupled to the memory; the memory is used for storing programs; the processor is configured to invoke a program stored in the memory to perform the method according to the first aspect embodiment and/or any possible implementation manner of the first aspect embodiment.

In a fourth aspect, embodiments of the present application further provide a storage medium, on which a computer program is stored, where the computer program is executed by a processor to perform the method provided in the foregoing first aspect and/or any one of the possible implementation manners of the first aspect.

Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the embodiments of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and drawings.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts. The foregoing and other objects, features and advantages of the application will be apparent from the accompanying drawings. Like reference numerals refer to like parts throughout the drawings. The drawings are not intended to be to scale as practical, emphasis instead being placed upon illustrating the subject matter of the present application.

Fig. 1 shows a schematic flow chart of a data processing method provided in an embodiment of the present application.

Fig. 2 is a schematic flow chart illustrating sales prediction according to an embodiment of the present disclosure.

Fig. 3 shows a schematic diagram of a classification principle for a product category to which an online product belongs according to an embodiment of the present application.

Fig. 4 shows a schematic diagram of a classification principle for a product category to which an offline product belongs according to an embodiment of the present application.

Fig. 5 is a schematic diagram illustrating a process of building a sales prediction model according to an embodiment of the present application.

Fig. 6 shows a block diagram of a data processing apparatus according to an embodiment of the present application.

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

Detailed Description

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. Meanwhile, relational terms such as "first," "second," and the like may be used solely in the description herein to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.

Further, the term "and/or" in the present application is only one kind of association relationship describing the associated object, and means that three kinds of relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone.

In view of the defects of poor applicability, poor prediction accuracy and the like of most of the existing offline or online promotion schemes. The embodiment of the present application provides a data processing method, when an optimal sales promotion strategy that maximizes the profit of N products participating in sales promotion pricing needs to be predicted, a decision factor influencing the profit maximization is obtained, and a sales prediction result corresponding to the price is obtained according to the price of each product in the N products, so as to solve a sales promotion product list that maximizes the profit and the price of each product in the product list, which will be described below with reference to the steps included in fig. 1.

Step S101: and acquiring a decision factor influencing the profit maximization.

And when refined promotion pricing needs to be carried out on the online stores or the offline stores, and the optimal promotion strategy of N products participating in promotion pricing and enabling the profit to be maximized is predicted, obtaining a decision factor influencing the profit maximization.

In an embodiment, a decision factor affecting profit maximization may be obtained according to configuration operations of a user, that is, the decision factor may be configured according to needs of the user. The configuration may be one-time configuration for multiple uses, that is, the decision factor may be configured in advance, and then, when a sales promotion pricing strategy is predicted, the decision factor influencing the profit maximization is directly obtained based on the previous configuration. Of course, the configuration may be real-time, that is, one configuration can only be used once, and whenever the promotion pricing strategy needs to be predicted, the currently acquired decision factors influencing the profit maximization are all acquired in response to the current configuration operation of the user and are real-time.

Wherein the decision factor comprises: decision objectives, decision variables, and constraints. The decision target is Return On Investment (ROI) or Gross capitalization (GMV). The decision variables include: the number of the products in the product list is not more than N, and N is an integer more than 1. The limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision-making target. For example, the user may specify a product list participating in promotion, prices of the N products, upper and lower price limits of the products in the product list, and a lower limit of the decision target, so as to predict the product list that maximizes the profit and prices of the products in the product list.

When the decision target is the return on investment ROI, the decision variables may further include: advertising inputs for each of the N products; accordingly, the limitation condition further includes: the advertising exposure fluctuation range of each product in the product list. That is, when the decision target is the return on investment ROI, in one embodiment, the decision variables include: a list of products participating in the promotion, respective prices of the N products, and respective advertising impressions of the N products; the limiting conditions include: the price fluctuation range of each product in the product list, the lower limit of the decision target, and the advertisement investment fluctuation range of each product in the product list.

It should be noted that the number of products in the product list is smaller than the number N of products participating in the promotion pricing, for example, if 15 products participate in the promotion pricing, the number of products in the product list participating in the promotion is specified to be smaller than or equal to 15, for example, 10 products participate in the promotion, that is, only 10 products participate in the promotion.

Step S102: and obtaining a sales prediction result corresponding to the price according to the price of each product in the N products.

And after the decision factor influencing the profit maximization is obtained, obtaining a sales forecasting result corresponding to the price according to the price of each product in the N products. As one implementation, the sales prediction results corresponding to the respective prices of the N products participating in the promotion pricing can be called through an interface function. When the decision factor further includes the parameter factor of advertisement input, correspondingly, according to the price of each product in the N products, the sales forecast result corresponding to the price is obtained, and the sales forecast result corresponding to the price and the advertisement input is obtained according to the price of each product in the N products and the advertisement input. That is, before that, the predicted sales of each of the N products participating in the pricing for promotion at different prices and different advertising impressions, which correspond to different product sales, needs to be obtained. The sales prediction result corresponding to each of the N products participating in the promotional pricing can be obtained through the sales prediction process shown in fig. 2. This process will be described below in conjunction with the steps involved in fig. 2.

Step S201: raw data relating to sales volume of each of the N products is obtained.

The method comprises the steps of obtaining original data of N products participating in promotion pricing, wherein the original data of the N products related to sales volume respectively, wherein the original data corresponding to the online products and the offline products can be the same or different, and for example, the original data can respectively comprise four types of data of product data representing product details (namely product basic information comprises information such as package, price, whether advertisement is invested or not, advertisement investment amount and the like), transaction data (comprising transaction records of past period), competitive product data (comprising basic information of competitive products) and other data. And when the product involved in the promotion pricing is an online product, other data now includes: at least one of behavior data including browsing and returning behavior of the product purchaser with respect to the product, and the shopping cart data in which the product is added to the shopping cart. When the product participating in the promotional pricing is an off-line product, other data at this time includes: store data of the store to which the product belongs (including basic information of the store (such as the size of the store, the location of the store, sales volume, and sales volume)). In addition, when predicting the sales volume of the product, whether the product is on-line or off-line, external data including weather conditions, holiday conditions and the like can be acquired, so that the accuracy of the sales volume prediction is improved.

In addition, the type of the raw data is not absolute, and the type may be increased or decreased according to need, for example, the data of the competitive products may be removed. Similarly, the data in other data may be increased or decreased accordingly, and thus, the original data in the above example is not to be construed as limiting the present application.

Step S202: and for each product, determining the category of the product and acquiring the input features according to the original data of the product.

After respective raw data of N products participating in promotion pricing are obtained, for each product, the category of the product is determined according to the raw data of the product, and input features of a sales prediction model are obtained.

When the category to which the product belongs is determined, each product can be clustered based on respective original data of the N products, and the category to which each product belongs can be obtained.

When clustering is performed, the products can be divided according to two dimensions of the online products from the perspective of the products and the perspective of the users, and the process can be shown in fig. 3. When the product is divided from the product perspective, the product is divided mainly according to the product data, the transaction data and the competitive product data, and the product can be divided according to whether the product is a seasonal commodity (a commodity greatly influenced by seasons is the seasonal commodity), whether the product is a price sensitive commodity (a commodity greatly influenced by price change on sales is the price sensitive commodity) and whether competitive commodities exist. When the division is performed from the perspective of the user, the division is performed mainly according to transaction data, basic information (consumer information) of a product purchaser, shopping cart data and behavior data, and the division can be performed from the perspective of population attributes, consumer behaviors, RFM scores (RFM describes the value condition of the consumer according to three indexes of recent purchase time (Recence, R), purchase Frequency (F) and purchase amount (Monetary, M)) of the consumer and the like. The population attribute classification is mainly performed according to static attributes such as age and occupation of consumers. The behavior division is mainly divided according to the characteristics of purchasing behaviors such as types of products frequently purchased by consumers. The RFM score division is mainly to calculate the user score according to three dimensions of the time length of the latest purchase of the user, the purchase frequency and the purchase amount, the RFM score is mainly used for measuring the purchase value of the user, and the RFM score division is carried out according to the size of the RFM score.

When clustering is performed, the products can be divided according to two dimensions of the offline products from the product perspective and the store perspective, and the process can be seen in fig. 4. When the product is divided from the product perspective, the product is divided mainly according to the product data, the transaction data and the competitive product data, and the product can be divided according to whether the product is a seasonal commodity (a commodity greatly influenced by seasons is the seasonal commodity), whether the product is a price sensitive commodity (a commodity greatly influenced by price change on sales is the price sensitive commodity) and whether competitive commodities exist. When the division is performed from the store perspective, the division is performed mainly based on store data and transaction data of the store to which the product belongs, and the division may be performed from the perspective of price sensitivity, store level, store radiation population attribute, consumer level around the store, and the like.

The price sensitivity classification mainly classifies stores into different price sensitivity levels according to sales promotion activities at the store level and the influence of prices on sales volume. The store level division mainly divides stores according to store level data such as the size, sales volume, and sales volume of the stores. The consumption level division around the store mainly divides the consumption levels of people around the store according to some statistical characteristics of the sales volume or the turnover of the store, such as the mean value, the variance and the like of the sales volume or the turnover. The attribute division of the radiation crowd of the store is mainly characterized in that the purchase attribute of the radiation crowd of the store is indirectly depicted according to the attribute of hot sales commodities of the store, the statistical characteristics of sales data and the performance of sales/sales on holidays of different festivals. For example, if a large part of hot-sold goods in a store belong to living goods (vegetables, kitchen supplies, etc.) and the peak sales of such goods is on the weekend, a large part of consumers in the store are presumably office workers.

When the input features of the sales prediction model are obtained according to the original data of the product, in order to make the predicted sales result as accurate as possible, the input features of the sales prediction model input thereto may include data of different dimensions such as sales data, price, and advertisement investment information. The sales data may be a sequence of daily average sales data of the past week, the prices may be a sequence of prices including the past week, and the advertisement insertion information may include information on whether or not an advertisement is inserted, the amount of money inserted, and the like. In addition, the input features may also include product information, such as information on the type of product (e.g., fruit, clothing, home appliance), size, packaging, etc.

For example, if there is no data of the dimension of the competitive product data in the raw data, when the category to which the product belongs is classified from the product perspective, the product is classified only based on whether the product is a seasonal commodity (a commodity that is greatly affected by the season is a seasonal commodity) or whether the product is a price sensitive commodity. The division of the above examples is therefore not to be construed as a limitation of the present application.

Step S203: and predicting the sales volume of the product by using the sales volume prediction model corresponding to the category and the input characteristics.

After the category of each product is determined according to the original data of the product and the input features are obtained, the sales of the product can be predicted by using the sales prediction model corresponding to the category and the obtained input features. Due to the fact that sales volume prediction results corresponding to different prices and different advertisement investments are different, the sales volumes of the different prices and the different advertisement investments need to be predicted so as to be called later. The method shown in fig. 2 can be used to obtain sales prediction results of different products at different prices and different advertisement investments, and the predicted sales prediction results are stored in a database, so that the sales prediction results corresponding to the prices of the N products can be obtained according to the prices of the respective products in the N products when the optimal sales promotion strategies of the N products participating in the sales promotion pricing need to be predicted.

In the embodiment of the application, sales prediction models are respectively constructed for different types of commodities, so that the sales of the commodities can be predicted more accurately. The sales prediction model is a sales prediction model trained in advance, that is, before step S203, the sales prediction model needs to be created and trained. The process of establishing the sales prediction model can be described with reference to the steps included in fig. 5.

Step S301: raw data relating to sales for each of the M products is obtained.

The sales volume prediction model is established, and sample data needs to be acquired, so in the embodiment of the application, original data of M products related to sales volume are acquired, wherein M is an integer and is greater than N. In order to make the established sales prediction model more accurate, the required raw data should be as much as possible, for example, M is more than one hundred.

The dimensions included in the raw data of each product should be the same as much as possible, and the raw data may be the same or different for the online product and the offline product, and the specific details thereof may refer to the contents recorded in step S201, and are not described herein for the sake of avoiding redundancy.

Step S302: and clustering the M products based on the respective original data of the M products to obtain a plurality of clustering categories.

After the original data of the M products related to sales volume are obtained, clustering is carried out on the M products based on the original data of the M products, and a plurality of clustering categories can be obtained, wherein each clustering category comprises at least one product. When clustering is performed, products can be classified from two dimensions of a product perspective and a user perspective for online products, and products can be classified from two dimensions of a product perspective and a store perspective for offline products, and specific details thereof can refer to the contents described in step S202, and will not be described here again to avoid redundancy.

Step S303: and respectively constructing a sales prediction model for each cluster category.

After M products are subdivided into different categories, a sales prediction model is respectively constructed for each clustering category. When the sales prediction model is constructed, the original data of a plurality of products belonging to the same clustering category are utilized, respective input features are extracted from the original data, the extracted input features are used as training samples, and the initial model is trained, so that the trained sales prediction model can be obtained, wherein the process of the training model is not different from that of the existing training model, which is well known by the technical personnel in the field and is not described here.

When the Model is selected, when the data size is not large, a bayesian causal inference Model and a multinomial Logit Model (MNL) Model may be selected. The MNL model constructs the probability of a user purchasing a certain product based on the utility function. The Bayesian model is different from a general statistical method, not only utilizes model information and data information, but also fully utilizes prior information, combines objective factors and subjective factors, and has more flexibility for the occurrence of abnormal conditions. When the data amount is large, various machine learning models or neural networks can be selected. Such as random forest, XGBOOST, etc., based on tree models, and may also select Neural network models, such as Deep Neural Networks (DNN), Long-Short Term Memory Networks (LSTM), etc.

In the method, the model is corrected by using the change of data distribution or data relation captured by online incremental learning when the external environment changes, and the model can predict the sales volume of a certain product under different conditions according to different input information.

Step S103: and solving an output result for maximizing the profit according to the sales prediction result corresponding to the price of each of the N products, the decision factor and the optimization algorithm.

After the decision factor and the sales prediction results corresponding to the respective prices of the N products are obtained, an output result for maximizing the profit can be solved according to the sales prediction results corresponding to the respective prices of the N products, the decision factor and the optimization algorithm, wherein the output result comprises: the product list and the prices of the products in the product list realize the product selection and pricing of fine promotion.

Wherein, the total volume of transaction = the price of product 1 + the sales volume corresponding to the price + the price of product 2 + the sales volume corresponding to the price + … + the price of product N + the sales volume corresponding to the price. The optimization algorithm has the function of calculating different prices and total volume of deals corresponding to the different sales according to the decision factors, and selecting a product list corresponding to the maximized total volume of deals and the price of each product in the product list as output.

When the decision target is the return on investment ROI, the decision variables may further include: advertising inputs for each of the N products; accordingly, the limitation condition further includes: the advertisement input fluctuation range of each product in the product list, and the output result further comprises: advertising impressions for each product in the product list. That is, the decision variables at this time include: a list of products participating in the promotion, respective prices of the N products, and respective advertising impressions of the N products; the limiting conditions include: the price fluctuation range of each product in the product list, the lower limit of a decision target and the advertisement input fluctuation range of each product in the product list; the output result is a product list, prices of products in the product list, and advertising investments of the products in the product list. The sales prediction result at this time is the price of each product in the N products and the sales prediction result corresponding to the advertisement investment, that is, the sales prediction result corresponding to the price and the advertisement investment is obtained according to the price and the advertisement investment of each product in the N products, after the decision factor, the price of each product in the N products and the sales prediction result corresponding to the advertisement investment are obtained, the output result for maximizing the profit can be solved according to the price of each product in the N products and the sales prediction result corresponding to the advertisement investment, the decision factor and the optimization algorithm, and at this time, the corresponding output result is the product list, the price of each product in the product list and the advertisement investment of each product in the product list. It should be noted that the output result depends on the decision variable.

The return on investment ROI = (total amount of transaction-total cost)/total cost, = (cost of product 1 + advertisement input) + (cost of product 2 + advertisement input) + … … + (cost of product N + advertisement input) + other costs (e.g. store front fee, water and electricity fee). The optimization algorithm has the effects that the investment return rates corresponding to different prices, different advertisement investments and different sales volumes are calculated according to decision factors, and a product list corresponding to the maximized investment return rate and the prices and the advertisement investments of the products in the product list are selected as output.

The optimization algorithm used in the solution may use various algorithms, such as nonlinear optimization, combinatorial optimization, and the like.

According to the data processing method provided by the embodiment of the application, on one hand, commodities are subdivided from multiple dimensions, and sales prediction models are respectively constructed for different types of product subdivision, so that the model precision is improved. On the other hand, the machine learning model is used for replacing the prior knowledge model, so that the early-stage market research work can be reduced to the greatest extent, and the problems that the prior model is over-assumed and the actual landing time deviation is large are solved. Meanwhile, when the sales volume is predicted, the data such as price and advertisement investment are taken into consideration, so that different sales volumes can be predicted according to different prices and advertisement investments, and then a sales promotion product list which can reach the maximum profit and the price of each product are solved according to the prediction results of the sales volumes of the products under different prices and different advertisement investments and decision factors which influence the maximization of the profit, so that the product selection and pricing for the fine promotion of the online products or the offline products are realized.

The embodiment of the present application further provides a data processing apparatus 100, as shown in fig. 6. The data processing apparatus 100 includes: an acquisition module 110, a calling module 120, and a solving module 130.

An obtaining module 110, configured to obtain a decision factor influencing revenue maximization when an optimal promotion strategy that maximizes revenue by predicting N products participating in promotion pricing is needed, where the decision factor includes: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target. Optionally, the obtaining module 110 is configured to obtain a decision factor influencing profit maximization in response to a configuration operation of a user.

And the calling module 120 is configured to obtain a sales prediction result corresponding to the price according to the price of each of the N products.

A solving module 130, configured to solve an output result that maximizes the profit according to the sales prediction result, the decision factor, and the optimization algorithm corresponding to the price of each of the N products, where the output result includes: a product list and prices of each product in the product list.

Optionally, when the decision target is a return on investment ROI, the decision variables further include: the advertisement input of each of the N products, and accordingly, the limitation further includes: the advertising input fluctuation range of each product in the product list, and the output result further includes: advertising impressions for each product in the product list. Correspondingly, the calling module 120 is further configured to obtain a sales prediction result corresponding to the price and the advertisement investment according to the price and the advertisement investment of each of the N products. Correspondingly, the solving module 130 is configured to solve the output result for maximizing the profit according to the price of each of the N products, the sales prediction result corresponding to the advertisement investment, the decision factor, and the optimization algorithm.

Wherein, this data processing apparatus further includes: and a prediction module. The prediction module is to: before the calling module 120 obtains the sales prediction result corresponding to the price according to the price of each of the N products, obtaining the original data of each of the N products related to the sales; aiming at each product, determining the category of the product and acquiring input characteristics according to the original data of the product, wherein the input characteristics comprise sales data, price and advertisement input information; and predicting the sales volume of the product by using the sales volume prediction model corresponding to the category and the input characteristics. Optionally, the prediction module is configured to cluster each product based on the respective raw data of the N products to obtain a category to which each product belongs. Optionally, the measurement module is configured to obtain product data, transaction data, auction data, and other data, which are related to sales volumes of the N products participating in the promotion pricing, where the other data includes, when the N products are all online products: basic information of a product purchaser, shopping cart data of a shopping cart into which the product is added, and behavior data including browsing and returning behaviors of the product purchaser for the product; when the N products are offline products, the other data includes: store data of the store to which the product belongs.

Wherein, this data processing apparatus further includes: and a model building module. The model building module is used for: before a prediction module predicts the sales volume of the product by using a sales volume prediction model corresponding to the category and the input characteristics, acquiring original data of M products related to the sales volume, wherein M is an integer and is larger than N; clustering the M products based on respective original data of the M products to obtain a plurality of clustering categories; and respectively constructing a sales prediction model for each cluster category.

The data processing apparatus 100 according to the embodiment of the present application has the same implementation principle and the same technical effect as those of the foregoing method embodiments, and for the sake of brief description, reference may be made to the corresponding contents in the foregoing method embodiments for the parts of the apparatus embodiments that are not mentioned.

As shown in fig. 7, fig. 7 is a block diagram illustrating a structure of an electronic device 200 according to an embodiment of the present disclosure. The electronic device 200 includes: a transceiver 210, a memory 220, a communication bus 230, and a processor 240.

The elements of the transceiver 210, the memory 220, and the processor 240 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the components may be electrically coupled to each other via one or more communication buses 230 or signal lines. The transceiver 210 is used for transceiving data. The memory 220 is used for storing a computer program, such as the software functional module shown in fig. 6, i.e., the data processing apparatus 100. The data processing apparatus 100 includes at least one software functional module, which may be stored in the memory 220 in the form of software or firmware (firmware) or solidified in an Operating System (OS) of the electronic device 200. The processor 240 is configured to execute executable modules stored in the memory 220, such as software functional modules or computer programs included in the data processing apparatus 100. For example, the processor 240 is configured to obtain a decision factor influencing profit maximization when an optimal promotion strategy is needed to predict that N products participating in promotion pricing maximize profit, wherein the decision factor includes: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target; and the price forecasting device is also used for obtaining the sales forecasting result corresponding to the price according to the price of each product in the N products; and the system is further used for solving an output result for maximizing the profit according to the sales prediction result corresponding to the price of each of the N products, the decision factor and the optimization algorithm, wherein the output result comprises: a product list and prices of each product in the product list.

The Memory 220 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.

The processor 240 may be an integrated circuit chip having signal processing capabilities. The processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor 240 may be any conventional processor or the like.

The electronic device 200 includes, but is not limited to, a web server, a database server, a cloud server, and the like.

The present embodiment also provides a non-volatile computer-readable storage medium (hereinafter, referred to as a storage medium), where the storage medium stores a computer program, and when the computer program is run by the electronic device 200, the computer program executes the steps included in the data processing method according to the foregoing method embodiment.

It should be noted that, in the present specification, the embodiments are all described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments may be referred to each other.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a notebook computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A data processing method, comprising:

when an optimal promotion strategy that N products participating in promotion pricing are predicted to maximize benefits is needed, obtaining a decision factor influencing the benefits maximization, wherein the decision factor comprises: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target;

obtaining a sales forecasting result corresponding to the price according to the price of each product in the N products;

solving an output result for maximizing the profit according to the sales forecasting result, the decision factor and the optimization algorithm corresponding to the price of each of the N products, wherein the output result comprises: a product list and prices of each product in the product list.

2. The method of claim 1, wherein when the decision target is a return on investment ROI, the decision variables further comprise: the advertisement input of each of the N products, and accordingly, the limitation further includes: the advertising input fluctuation range of each product in the product list, and the output result further includes: advertising investment for each product in the product list; accordingly, the number of the first and second electrodes,

obtaining a sales prediction result corresponding to the price according to the price of each product in the N products, wherein the sales prediction result comprises the following steps:

obtaining a sales prediction result corresponding to the price and the advertisement investment according to the price and the advertisement investment of each product in the N products; accordingly, the number of the first and second electrodes,

solving an output result for maximizing the profit according to the sales forecasting result corresponding to the price of each of the N products, the decision factor and the optimization algorithm, wherein the output result comprises the following steps:

and solving an output result for maximizing the profit according to the price of each of the N products, the sales prediction result corresponding to the advertisement investment, the decision factor and the optimization algorithm.

3. The method of claim 1, wherein obtaining a decision factor that affects revenue maximization comprises:

and responding to the configuration operation of the user to obtain a decision factor influencing the profit maximization.

4. The method of claim 1, wherein before obtaining a sales prediction corresponding to a price of each of the N products based on the price, the method further comprises:

acquiring original data of the N products related to sales volume;

aiming at each product, determining the category of the product and acquiring input characteristics according to the original data of the product, wherein the input characteristics comprise sales data, price and advertisement input information;

and predicting the sales volume of the product by using the sales volume prediction model corresponding to the category and the input characteristics.

5. The method of claim 4, wherein determining the category to which the product belongs based on the raw data of the product comprises:

and clustering each product based on the respective original data of the N products to obtain the category of each product.

6. The method of claim 4, wherein obtaining raw data relating to sales for each of the N products participating in promotional pricing comprises:

obtaining product data, transaction data, auction data, and other data relating to sales for each of the N products participating in promotional pricing, wherein the other data comprises, when the N products are all online products: basic information of a product purchaser, shopping cart data of a shopping cart into which the product is added, and behavior data including browsing and returning behaviors of the product purchaser for the product; when the N products are offline products, the other data includes: store data of the store to which the product belongs.

7. The method of claim 4, wherein prior to predicting sales of the product using the sales prediction model for the category and the input features, the method further comprises:

acquiring original data of M products related to sales volume respectively, wherein M is an integer and is larger than N;

clustering the M products based on respective original data of the M products to obtain a plurality of clustering categories;

and respectively constructing a sales prediction model for each cluster category.

8. A data processing apparatus, comprising:

an obtaining module, configured to obtain a decision factor influencing revenue maximization when an optimal promotion strategy that maximizes revenue by predicting N products participating in promotion pricing is needed, where the decision factor includes: decision objectives, decision variables and constraints; the decision target is an investment return rate ROI or a total volume of trades GMV; the decision variables include: a product list participating in promotion and respective prices of the N products, wherein the number of the products in the product list is not more than N, and N is an integer more than 1; the limiting conditions include: the price fluctuation range of each product in the product list and the lower limit of the decision target;

the calling module is used for acquiring a sales forecasting result corresponding to the price according to the price of each product in the N products;

a solving module, configured to solve an output result that maximizes a benefit according to the sales prediction result, the decision factor, and the optimization algorithm corresponding to the price of each of the N products, where the output result includes: a product list and prices of each product in the product list.

9. An electronic device, comprising:

a memory and a processor, the processor coupled to the memory;

the memory is used for storing programs;

the processor to invoke a program stored in the memory to perform the method of any of claims 1-7.

10. A storage medium having stored thereon a computer program which, when executed by a processor, performs the method according to any one of claims 1-7.