Commodity information processing method, device, computer equipment and storage medium

Abstract

The present invention provides a product information processing method, apparatus, computer device, and storage medium. The method comprises obtaining online and offline product data; importing the online and offline data into a preset frequent item set mining algorithm to obtain frequent item sets for the online and offline data; calculating association rules for each product based on the frequent item sets based on a preset association rule algorithm; importing the online and offline data into a product recommendation model, calculating product recommendation information using the product recommendation model; and generating a product recommendation scheme based on the association rules and product recommendation information. By combining the online and offline data to generate product association rules and product recommendation information, offline product inventory can also be managed based on online attention levels, reducing inventory pressure, making product recommendations more accurate, greatly optimizing the user's purchasing experience, and effectively improving product sales efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of commodity information processing, and in particular to a commodity information processing method, apparatus, computer equipment and storage medium. Background Art

[0002] At present, with the development of the Internet, more and more people choose to buy goods online. Although online shopping provides us with great convenience, since users cannot see the real thing and cannot further understand the product information, users buy goods without sufficient understanding of the products. Although offline stores allow customers to directly view and experience, offline stores lack understanding of user needs, resulting in the inability of offline stores' goods to match user needs, causing users to often be unable to buy suitable goods in offline stores. In the long run, this also causes increased inventory pressure on offline stores. Summary of the Invention

[0003] Based on this, it is necessary to provide a commodity information processing method, device, computer equipment and storage medium to address the above technical problems.

[0004] A method for processing commodity information, comprising:

[0005] Obtain online and offline data of products;

[0006] Importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0007] Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets;

[0008] Importing the online data and the offline data into a pre-built product recommendation model, and calculating product recommendation information using the product recommendation model;

[0009] Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

[0010] In one embodiment, the step of importing the online data and the offline data into a pre-built product recommendation model and calculating product recommendation information using the product recommendation model includes:

[0011] Preprocessing the online data and the offline data to obtain preprocessed data;

[0012] Importing the preprocessed data into a neural network for learning, and constructing a plurality of active neurons based on the cross-combination of the preprocessed data;

[0013] Detecting whether the error of each active neuron is less than a preset error threshold;

[0014] When the error of the active neuron is less than the preset error threshold, the active neuron is eliminated; when the error of the active neuron is greater than or equal to the preset error threshold, the active neuron is merged;

[0015] The product recommendation model is constructed based on the fused multiple active neurons.

[0016] In one embodiment, the step of importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data includes:

[0017] Classifying the online data and the offline data according to preset classification rules to obtain classified data;

[0018] The classified data is imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classified data.

[0019] In one embodiment, the step of importing the classified data into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classified data includes:

[0020] Counting the number of occurrences of each of the classification data;

[0021] Based on the number of occurrences of each of the classification data, the proportion of each of the classification data to the total classification data is calculated to obtain a data proportion;

[0022] Filtering out the classified data whose data ratio is greater than a preset ratio to obtain filtered data;

[0023] The filtered data are imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of each filtered data.

[0024] In one embodiment, the step of calculating the association rules of the commodities according to the frequent item sets based on a preset association rule algorithm includes:

[0025] Calculating the imbalance factor of each of the commodities using an imbalance factor algorithm;

[0026] Eliminating the commodities whose imbalance factors are greater than a preset factor;

[0027] Based on a preset association rule algorithm, the association rules of the commodities after elimination are calculated according to the frequent item sets.

[0028] In one embodiment, the preset frequent item set mining algorithm includes an FP Tree algorithm and an Apriori algorithm.

[0029] In one embodiment, the offline data of the product is obtained through a preset sensor.

[0030] A commodity information processing device, comprising:

[0031] Product data acquisition module, used to obtain online and offline data of products;

[0032] A frequent itemset calculation module is used to import the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0033] An association rule obtaining module, configured to obtain an association rule for each of the commodities based on the frequent item sets using a preset association rule algorithm;

[0034] A recommendation information obtaining module, configured to import the online data and the offline data into a pre-built product recommendation model, and calculate product recommendation information using the product recommendation model;

[0035] The product recommendation scheme generating module is used to generate a product recommendation scheme based on the association rules and the product recommendation information, and output the product recommendation scheme.

[0036] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and wherein the processor performs the following steps when executing the computer program:

[0037] Obtain online and offline data of products;

[0038] Importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0039] Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets;

[0040] Importing the online data and the offline data into a pre-built product recommendation model, and calculating product recommendation information using the product recommendation model;

[0041] Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

[0042] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0043] Obtain online and offline data of products;

[0044] Importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0045] Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets;

[0046] Importing the online data and the offline data into a pre-built product recommendation model, and calculating product recommendation information using the product recommendation model;

[0047] Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

[0048] The above-mentioned product information processing method, device, computer equipment and storage medium combine online data and offline data to generate product association rules and product recommendation information, so that products sold online can be recommended according to the level of attention of offline users, and offline products can also be managed in inventory according to the level of online attention, reducing inventory pressure, making product recommendations more accurate, greatly optimizing the user's experience of purchasing products, and effectively improving the sales efficiency of products. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 1 is a flow chart of a method for processing commodity information in one embodiment;

[0050] Figure 2 is a structural block diagram of a commodity information processing device in one embodiment;

[0051] Figure 3 is a diagram of the internal structure of a computer device in one embodiment;

[0052] Figure 4A A schematic diagram of the process of generating association rules for commodities in one embodiment;

[0053] Figure 4B A schematic diagram of the process of constructing a product recommendation model for products in one embodiment;

[0054] Figure 4C A schematic diagram of the process of generating a product recommendation solution for a product in an embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0056] Example 1

[0057] In this embodiment, Figure 1 As shown, a commodity information processing method is provided, which includes:

[0058] Step 110: Obtain online and offline data of the product.

[0059] In this embodiment, the online data is the data corresponding to the goods sold on the online network platform. The online data includes at least browsing data, click data, sales data and positioning data. For example, the browsing data is the number of times the product page is browsed by users and the browsing time of a single user. The click data is the number of times the link of the product is clicked. The sales data is the sales volume of the product. The positioning data is the location of the user who browses the product page or purchases the product.

[0060] Offline data refers to data corresponding to products sold offline. It includes at least attention data and purchase data. Attention data refers to data on users' attention to a particular product. This attention data may include the average length of time a user spends in front of a product and the number of times a user visits the product. This attention data reflects the user's level of interest in the product. Purchase data refers to the number of times a user purchases the product.

[0061] In this embodiment, the online data of the product can be obtained by collecting data from a network platform, and the offline data can be obtained by sensors. In one embodiment, the offline data of the product is obtained by a preset sensor. The preset sensor may include an infrared sensor, an iris sensor, and a photosensitive element. The sensor may include a camera. For example, the infrared sensor is used to detect users stopping in front of or passing by the product. The iris sensor is used to detect whether the user who repeatedly passes by the product is the same user and to count the number of users' passes. The camera is used to capture the user's facial image or to detect the number of users who stop by or pass by the product.

[0062] Step 120: Import the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data.

[0063] In this embodiment, the preset frequent itemset mining algorithm is used to calculate and obtain the frequent itemset of each commodity. The frequent itemset is a set of commodities whose support is greater than or equal to the minimum support.

[0064] In one embodiment, the preset frequent itemset mining algorithm includes an FP Tree algorithm and an Apriori algorithm.

[0065] In this embodiment, the preset frequent item set mining algorithm is further explained as the FP (Frequent Pattern) Tree algorithm. The FP Tree algorithm is also called FP-growth or FP growth algorithm. The online data and offline data are respectively constructed using the FP Tree algorithm to construct the FP tree of product information, and then the FP tree is used to find the conditional pattern to construct a conditional FP tree, and finally obtain the frequent item set.

[0066] In one embodiment, the step of importing the online data and the offline data into a preset frequent item set mining algorithm to obtain frequent item sets of the online data and the offline data includes: classifying the online data and the offline data according to preset classification rules to obtain classified data; importing the classified data into a preset frequent item set mining algorithm for calculation to obtain frequent item sets of the classified data.

[0067] Specifically, the FP Tree algorithm is used to construct the FP tree of the product information for the classified data, and then the FP tree is used to find the conditional pattern to construct a conditional FP tree, and finally the frequent item set is obtained.

[0068] Step 130 : Based on a preset association rule algorithm, an association rule of each of the commodities is calculated according to the frequent item sets.

[0069] In one embodiment, the preset association rule algorithm is an algorithm based on the Kulc criterion.

[0070] In this embodiment, based on the different frequent item sets obtained in the previous step, an algorithm based on the Kulc criterion is used to calculate the association relationships between products, thereby obtaining product association rules. These association rules are used to reflect the association relationships between products. For example, if a user purchases product A, they will generally also purchase product B. Therefore, a purchase association relationship exists between products A and B. Therefore, as a product recommendation solution, after a user purchases product A, product B or similar products of product B can be recommended to the user.

[0071] Step 140: Import the online data and the offline data into a pre-built product recommendation model, and calculate product recommendation information through the product recommendation model.

[0072] In this embodiment, a large amount of online and offline data is imported into a pre-built product recommendation model to calculate product recommendation information, which is used to recommend products, and recommend products with high demand and popularity to users.

[0073] Step 150: Generate a product recommendation plan based on the association rules and the product recommendation information, and output the product recommendation plan.

[0074] In this embodiment, a product recommendation plan is generated by combining the product association rules and product recommendation information, and the product recommendation plan is output to the online network platform and offline stores, so that the online network platform can push products to users according to the product recommendation plan, and the offline stores can manage inventory according to the product recommendation plan to reduce inventory backlogs, and the shopping guide screen can display the information of recommended products, thereby greatly optimizing the user's experience of purchasing products and effectively improving the sales efficiency of products.

[0075] In the above embodiment, online data and offline data are combined to generate product association rules and product recommendation information, so that products sold online can be recommended based on the level of attention of offline users, and offline products can also be managed in inventory based on the level of online attention, reducing inventory pressure and making product recommendations more accurate, greatly optimizing the user's experience in purchasing products and effectively improving product sales efficiency.

[0076] In one embodiment, the step of importing the online data and the offline data into a pre-built product recommendation model and calculating the product recommendation information through the product recommendation model includes: pre-processing the online data and the offline data to obtain pre-processed data; importing the pre-processed data into a neural network for learning, and constructing a plurality of active neurons based on the cross-combination of the pre-processed data; detecting whether the error of each active neuron is less than a preset error threshold; when the error of the active neuron is less than the preset error threshold, eliminating the active neuron; when the error of the active neuron is greater than or equal to the preset error threshold, fusing the active neurons; and constructing the product recommendation model based on the fused plurality of active neurons.

[0077] In this embodiment, based on a large amount of historical data, historical online data and offline data are preprocessed. The preprocessed online data includes browsed product information, browsing time, whether to add to the shopping cart, mobile phone positioning, historical purchase records, and product evaluation information; the preprocessed offline data includes browsed product information, browsing time, user satisfaction with the product, product evaluation, number of people entering the store every day, and purchase records.

[0078] In this embodiment, the product recommendation model is implemented using the GMDH (Group Method of Data Handling) neural network learning algorithm. This algorithm uses preprocessed data as input and generates a series of active neurons by cross-combining them two by two. The errors and mean squared errors are calculated and compared with a preset error threshold. These active neurons are then eliminated and integrated. After multiple iterations, the optimal recommendation model, i.e., the product recommendation model, is ultimately constructed. This makes the product recommendation model more accurate.

[0079] In one embodiment, the step of importing the classified data into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classified data includes: counting the number of times each of the classified data appears; based on the number of times each of the classified data appears, calculating the proportion of each of the classified data to all the classified data to obtain the data proportion; filtering out the classified data whose data proportion is greater than the preset proportion to obtain the filtered data; importing the filtered data into the preset frequent item set mining algorithm for calculation to obtain the frequent item sets of each of the filtered data.

[0080] In this embodiment, the number of occurrences of each classified data is first calculated based on the classified data, referred to as the number of occurrences of the frequent set. According to the number of occurrences of each classified data, the data with a minimum support greater than a preset ratio is filtered out. In this embodiment, the preset ratio is 20%, and the minimum support is the ratio of the classified data to the total data. Finally, the selected data is constructed into an FP tree through the FP Tree algorithm, and the sorted frequent item sets can be obtained from the FP tree.

[0081] In one embodiment, the step of calculating the association rules of each of the commodities according to the frequent item sets based on a preset association rule algorithm includes: calculating the imbalance factor of each of the commodities using an imbalance factor algorithm; eliminating the commodities whose imbalance factors are greater than the preset factors; and calculating the association rules of each of the commodities after elimination according to the frequent item sets based on a preset association rule algorithm.

[0082] In this embodiment, the imbalance factor RI of the associated products is calculated by the imbalance factor algorithm, and data with RI>0.6 is filtered out, and the product association rules are finally obtained.

[0083] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0084] Example 2

[0085] This application uses machine learning to build models based on data mining from both online and offline data to provide product recommendations for both online and offline stores. This effectively increases the online exposure of popular offline store products, and vice versa, improving the user shopping experience and boosting the company's economic benefits.

[0086] The specific process is as follows:

[0087] 1. Commodity association analysis, such as Figure 4A As shown:

[0088] 1. Online, the browsing, exposure, and sales data of online products are obtained through programs; offline, sensors and shopping guide screens are used to collect product attention (popularity), offline store product sales information, and offline store location information.

[0089] 2. Filter invalid data based on the data collected online and offline, and classify the filtered data according to date, sales area, product category, and sales channel (online and offline).

[0090] 3. Use the FP Tree algorithm to construct an FP tree of product information for the classified data, and then use the FP tree to find the conditional pattern base to construct a conditional FP tree, and finally obtain a frequent item set (a set with support greater than or equal to the minimum support, which mainly records the frequency of occurrence of different data).

[0091] The specific process of obtaining frequent itemsets is as follows:

[0092] 1. Calculate the number of occurrences of each data item based on the classified data (referred to as the number of occurrences of the frequent set)

[0093] 2. From the data obtained in step 1, filter out the data with a minimum support (the proportion of data to the total data) greater than 20%.

[0094] 3. The data obtained in step 2 is used to construct an FP tree through the algorithm, and the sorted frequent item sets can be obtained from the FP tree.

[0095] 4. Based on the resulting frequent item sets, the Kulc criterion (implemented by the algorithm) is used to calculate the association relationships between products. The specific process involves using the imbalance factor algorithm to calculate the imbalance factor (RI) of the associated products. Data with an RI > 0.6 (this data can be manually adjusted in practice) is filtered out, ultimately resulting in the product association rules.

[0096] 2. Build a recommendation model, such as Figure 4B As shown:

[0097] 1. Based on a large amount of historical data, historical online and offline data are pre-processed. The processed online data includes browsed product information, browsing time, whether to add products to the shopping cart, mobile phone location (with user consent), historical purchase records, and product evaluation information; the processed offline data includes browsed product information, browsing time, user satisfaction with the products, product evaluations, daily store visits, and purchase records.

[0098] 2. The recommendation model is implemented using the GMDH neural network learning algorithm. The preprocessed data in step 1 is used as the input source. A series of active neurons are generated by cross-combining the input data in pairs. The mean square error (MSE) is calculated and compared with a specified threshold (the threshold is manually set and needs to be continuously adjusted during the construction of the recommendation model to select the optimal threshold. Generally, the threshold can be selected as 0.6 or 0.45). These neurons are eliminated and integrated. After continuous iteration, the optimal recommendation model is finally constructed.

[0099] 3. Generate product display and related product recommendation plans, such as Figure 4C As shown:

[0100] Based on product association rules, historical data and real-time data, a recommendation model is used to generate product display and related product recommendation plans. The plan consists of two parts, corresponding to online and offline plans respectively.

[0101] The details of the online program are as follows:

[0102] 1. Through the analysis of offline popular product data, products with high offline store attention and high user satisfaction in different regions and their related products are pushed to the online APP according to different regions.

[0103] 2. If the online app user agrees to obtain location information, in addition to recommending popular online products, the app will also recommend products provided in the plan. When viewing products, the user will be provided with the address of the nearest offline store where the product is displayed, as well as offline user reviews of the product.

[0104] The specific contents of the offline plan are as follows:

[0105] 1. By analyzing the data of popular online products, products with high online click-through rates, large orders, and high user satisfaction, as well as their related products, are pushed to the offline store smart shopping guide screens. At the same time, stores can manage warehouse inventory based on this solution to reduce inventory backlogs, and display the products in the solution in stores to improve the user shopping experience.

[0106] 2. When users browse products on the smart shopping guide screen in offline stores, the evaluation labels provided in the plan will be displayed to users in the form of subtitles or voice, so that users can buy with confidence; at the same time, after viewing the product (sample), users can evaluate the product on the shopping guide screen through labels or customized forms. In addition, after viewing the product, related products will be recommended to users for selection.

[0107] Example 3

[0108] In this embodiment, Figure 2 As shown, a commodity information processing device is provided, comprising:

[0109] The product data acquisition module 210 is used to acquire online and offline data of products;

[0110] A frequent itemset calculation module 220 is configured to import the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0111] An association rule obtaining module 230 is configured to obtain an association rule for each of the commodities based on the frequent item sets using a preset association rule algorithm;

[0112] A recommendation information obtaining module 240 is configured to import the online data and the offline data into a pre-built product recommendation model and calculate product recommendation information using the product recommendation model;

[0113] The product recommendation scheme generating module 250 is configured to generate a product recommendation scheme based on the association rules and the product recommendation information, and output the product recommendation scheme.

[0114] In one embodiment, the commodity information processing device further includes:

[0115] A preprocessing module, configured to preprocess the online data and the offline data to obtain preprocessed data;

[0116] A neuron generation module is used to import the pre-processed data into a neural network for learning, and to construct a plurality of active neurons based on the cross-combination of the pre-processed data;

[0117] an error detection module, configured to detect whether the error of each active neuron is less than a preset error threshold;

[0118] A neuron elimination and fusion module is configured to eliminate the active neuron when the error of the active neuron is less than the preset error threshold; and to fuse the active neuron when the error of the active neuron is greater than or equal to the preset error threshold;

[0119] The recommendation model building module is used to build the product recommendation model based on the fused multiple active neurons.

[0120] In one embodiment, the frequent item set calculation module includes:

[0121] a classification unit, configured to classify the online data and the offline data according to preset classification rules to obtain classified data;

[0122] The frequent item set obtaining unit is used to import the classification data into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classification data.

[0123] In one embodiment, the frequent item set obtaining unit is further used to count the number of times each of the classified data appears; based on the number of times each of the classified data appears, calculate the proportion of each of the classified data to all the classified data to obtain the data proportion; filter out the classified data whose data proportion is greater than a preset proportion to obtain filtered data; import the filtered data into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of each of the filtered data.

[0124] In one embodiment, the association rule obtaining module includes:

[0125] an imbalance factor obtaining unit, configured to calculate the imbalance factor of each of the commodities using an imbalance factor algorithm;

[0126] a rejection unit, configured to reject the commodity whose imbalance factor is greater than a preset factor;

[0127] The association rule obtaining unit is configured to obtain the association rule of each of the commodities after elimination by calculation according to the frequent item sets based on a preset association rule algorithm.

[0128] In one embodiment, the preset frequent itemset mining algorithm includes an FP Tree algorithm and an Apriori algorithm.

[0129] In one embodiment, the offline data of the commodity is obtained through a preset sensor.

[0130] The specific definition of the product information processing device can be found in the definition of the product information processing method above and will not be repeated here. Each unit in the above-mentioned product information processing device can be implemented in whole or in part through software, hardware, or a combination thereof. Each of the above-mentioned units can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each of the above-mentioned units.

[0131] Example 4

[0132] In this embodiment, a computer device is provided. Its internal structure diagram can be shown as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, display screen and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program, and the non-volatile storage medium is deployed with a database, which is used to store online and offline data of goods. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with other computer devices deployed with application software. When the computer program is executed by the processor, a method for processing product information is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0133] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0134] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0135] Obtain online and offline data of products;

[0136] Importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0137] Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets;

[0138] Importing the online data and the offline data into a pre-built product recommendation model, and calculating product recommendation information using the product recommendation model;

[0139] Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

[0140] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0141] Preprocessing the online data and the offline data to obtain preprocessed data;

[0142] Importing the preprocessed data into a neural network for learning, and constructing a plurality of active neurons based on the cross-combination of the preprocessed data;

[0143] Detecting whether the error of each active neuron is less than a preset error threshold;

[0144] When the error of the active neuron is less than the preset error threshold, the active neuron is eliminated; when the error of the active neuron is greater than or equal to the preset error threshold, the active neuron is merged;

[0145] The product recommendation model is constructed based on the fused multiple active neurons.

[0146] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0147] Classifying the online data and the offline data according to preset classification rules to obtain classified data;

[0148] The classified data is imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classified data.

[0149] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0150] Counting the number of occurrences of each of the classification data;

[0151] Based on the number of occurrences of each of the classification data, the proportion of each of the classification data to the total classification data is calculated to obtain a data proportion;

[0152] Filtering out the classified data whose data ratio is greater than a preset ratio to obtain filtered data;

[0153] The filtered data are imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of each filtered data.

[0154] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0155] Calculating the imbalance factor of each of the commodities using an imbalance factor algorithm;

[0156] Eliminating the commodities whose imbalance factors are greater than a preset factor;

[0157] Based on a preset association rule algorithm, the association rules of the commodities after elimination are calculated according to the frequent item sets.

[0158] In one embodiment, the preset frequent itemset mining algorithm includes an FP Tree algorithm and an Apriori algorithm.

[0159] In one embodiment, the offline data of the commodity is obtained through a preset sensor.

[0160] Example 5

[0161] In this embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0162] Obtain online and offline data of products;

[0163] Importing the online data and the offline data into a preset frequent itemset mining algorithm to obtain frequent itemsets of the online data and the offline data;

[0164] Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets;

[0165] Importing the online data and the offline data into a pre-built product recommendation model, and calculating product recommendation information using the product recommendation model;

[0166] Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

[0167] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0168] Preprocessing the online data and the offline data to obtain preprocessed data;

[0169] Importing the preprocessed data into a neural network for learning, and constructing a plurality of active neurons based on the cross-combination of the preprocessed data;

[0170] Detecting whether the error of each active neuron is less than a preset error threshold;

[0171] When the error of the active neuron is less than the preset error threshold, the active neuron is eliminated; when the error of the active neuron is greater than or equal to the preset error threshold, the active neuron is merged;

[0172] The product recommendation model is constructed based on the fused multiple active neurons.

[0173] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0174] Classifying the online data and the offline data according to preset classification rules to obtain classified data;

[0175] The classified data is imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of the classified data.

[0176] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0177] Counting the number of occurrences of each of the classification data;

[0178] Based on the number of occurrences of each of the classification data, the proportion of each of the classification data to the total classification data is calculated to obtain a data proportion;

[0179] Filtering out the classified data whose data ratio is greater than a preset ratio to obtain filtered data;

[0180] The filtered data are imported into a preset frequent item set mining algorithm for calculation to obtain the frequent item sets of each filtered data.

[0181] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0182] Calculating the imbalance factor of each of the commodities using an imbalance factor algorithm;

[0183] Eliminate the products whose imbalance factors are greater than a preset factor;

[0184] Based on a preset association rule algorithm, the association rules of the commodities after elimination are calculated according to the frequent item sets.

[0185] In one embodiment, the preset frequent itemset mining algorithm includes an FP Tree algorithm and an Apriori algorithm.

[0186] In one embodiment, the offline data of the commodity is obtained through a preset sensor.

[0187] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0188] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0189] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A commodity information processing method, characterized in that: include: Obtain online and offline data for products; wherein the online data includes browsing data, click data, sales data, and location data for products sold online; and the offline data includes attention data and purchase data for products sold in offline stores; Classifying the online data and the offline data according to a preset classification rule to obtain classified data; counting the number of occurrences of each of the classified data; calculating the proportion of each of the classified data to the total classified data based on the number of occurrences of each of the classified data to obtain a data proportion; filtering out the classified data having a data proportion greater than a preset proportion to obtain filtered data; importing the filtered data into a preset frequent item set mining algorithm for calculation to obtain frequent item sets of each of the filtered data; Based on a preset association rule algorithm, the association rule of each of the commodities is calculated according to the frequent item sets; Preprocessing the online data and the offline data to obtain preprocessed data; importing the preprocessed data into a neural network for learning, constructing a plurality of active neurons based on the preprocessed data by cross-combining two by two; detecting whether the error of each active neuron is less than a preset error threshold; when the error of the active neuron is less than the preset error threshold, eliminating the active neuron; when the error of the active neuron is greater than or equal to the preset error threshold, fusing the active neurons; and constructing a product recommendation model based on the fused plurality of active neurons; Importing the online data and the offline data into the pre-built product recommendation model, and calculating product recommendation information through the product recommendation model; Based on the association rules and the product recommendation information, a product recommendation plan is generated and outputted.

2. The method according to claim 1, characterized in that The step of calculating the association rules of the commodities according to the frequent item sets based on the preset association rule algorithm includes: Calculating the imbalance factor of each of the commodities using an imbalance factor algorithm; Eliminate the products whose imbalance factors are greater than a preset factor; Based on a preset association rule algorithm, the association rules of the commodities after elimination are calculated according to the frequent item sets.

3. The method according to claim 1, characterized in that The preset frequent item set mining algorithms include FP Tree algorithm and Apriori algorithm.

4. The method according to any one of claims 1 to 3, characterized in that The offline data of the product is obtained through a preset sensor.

5. A commodity information processing device, characterized in that: include: The product data acquisition module is used to acquire online and offline data of products; wherein the online data includes browsing data, click data, sales data and location data of products sold online; the offline data includes attention data and purchase data of products sold in offline stores; A frequent item set calculation module is used to classify the online data and the offline data according to a preset classification rule to obtain classified data; count the number of occurrences of each of the classified data; calculate the proportion of each of the classified data to the total classified data based on the number of occurrences of each of the classified data to obtain a data proportion; filter out the classified data with a data proportion greater than a preset proportion to obtain filtered data; import the filtered data into a preset frequent item set mining algorithm for calculation to obtain frequent item sets of each of the filtered data; An association rule obtaining module, configured to obtain an association rule for each of the commodities based on the frequent item sets using a preset association rule algorithm; A recommendation model construction module is configured to preprocess the online data and the offline data to obtain preprocessed data; import the preprocessed data into a neural network for learning, and construct a plurality of active neurons based on the preprocessed data by cross-combining two by two; detect whether the error of each active neuron is less than a preset error threshold; when the error of the active neuron is less than the preset error threshold, eliminate the active neuron; when the error of the active neuron is greater than or equal to the preset error threshold, fuse the active neurons; and construct a product recommendation model based on the fused plurality of active neurons; A recommendation information obtaining module, configured to import the online data and the offline data into the pre-built product recommendation model, and calculate product recommendation information using the product recommendation model; The product recommendation scheme generating module is used to generate a product recommendation scheme based on the association rules and the product recommendation information, and output the product recommendation scheme.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images