E-commerce user behavior analysis system based on big data
By using pre-trained multi-layer neural network and self-attention mechanism to analyze user behavior in the e-commerce user behavior analysis system, the problems of large computing resources and poor prediction accuracy of existing systems are solved, and the user behavior of e-commerce platform is realized Accurate prediction and personalized recommendations improve transaction rate and user experience.
Patent Information
- Application Number
- CN202311700405.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-12
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2043-12-12
AI Technical Summary
The existing e-commerce user behavior analysis system uses too much computing resources to mine user behavior information from huge user and product data, and the accuracy of the prediction results is quite different from user behavior and demand, resulting in an e-commerce platform. The effect of use is not ideal.
Provide an e-commerce user behavior analysis system based on big data, including information collection module, data analysis module and recommendation generation module. Through a pre-trained multi-layer neural network model, combined with a self-attention mechanism, it analyzes user information and behavior information. Generate behavior prediction information and personalize recommendations based on user personality type and product information.
It greatly reduces the use of computing resources, realizes accurate prediction of user behavior of e-commerce platforms, and improves user transaction rate and user experience.
Smart Images

Figure CN118096285B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of e-commerce artificial intelligence applications, and in particular to an e-commerce user behavior analysis system based on big data. Background Art
[0002] Currently, in e-commerce platforms, due to the limitation of the salesperson’s personal energy, a salesperson cannot receive multiple users at the same time. When multiple users consult the salesperson at the same time, it is very easy for the salesperson’s recommendation to go wrong, leading to customer loss.
[0003] In order to improve the sales efficiency of e-commerce platforms, some user behavior analysis systems are currently being applied. Data mining, machine learning and other technologies are mainly used to analyze user behavior data to discover user preferences, purchase intentions and other information, and provide corresponding optimization suggestions.
[0004] However, due to the diversity and complexity of user behavior, existing behavior analysis systems not only need to occupy too much computing resources to mine user behavior information from huge user and product data, but the accuracy of the predicted results also deviates greatly from user behavior and needs, resulting in less than ideal results for the use of behavior analysis systems on e-commerce platforms. Summary of the invention
[0005] The present application provides an e-commerce user behavior analysis system based on big data, which aims to solve the problem that the existing behavior analysis system not only takes up too much computing resources to mine user behavior information from huge user and product data, but also the accuracy of the predicted results deviates greatly from the user's behavior and needs, resulting in less than ideal use effect of the behavior analysis system on the e-commerce platform.
[0006] In a first aspect, the present application provides an e-commerce user behavior analysis system based on big data, the system comprising:
[0007] An information collection module, used to collect user information corresponding to a user terminal and behavior information corresponding to the user information;
[0008] A data analysis module, used for inputting the user information and the behavior information into a pre-trained behavior analysis model, wherein the behavior analysis model is used for generating behavior prediction information according to the user information and the behavior information;
[0009] The recommendation generation module is used to obtain the intended product information corresponding to the behavior prediction information, determine the recommendation method information according to the user information, and complete the recommendation to the user terminal according to the recommendation method information.
[0010] In some embodiments, the system also includes: an information confirmation module, used to confirm the missing data information, blank data information and white noise data information corresponding to the user information and the behavior information based on the user information and the behavior information; a data cleaning module, used to fill the missing data information, delete the blank data information, and denoise the white noise data information according to the data type corresponding to the missing data information, so as to complete the data cleaning of the user information and the behavior information.
[0011] In some embodiments, the behavior analysis model includes a preset multi-layer neural network, which includes at least a fully connected layer and a coding layer; the system also includes: an information collection unit, used to collect the user information and the product browsing information and product operation information of the user terminal; a data analysis unit, used to input the user information and the product browsing information and product operation information of the user terminal into the fully connected layer for full connection, and input the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer outputs the behavior prediction information based on the self-attention mechanism.
[0012] In some embodiments, the coding layer stores an attention function, and the attention function is used to calculate the first attention parameter of the product browsing information and the user information, and the second attention parameter of the product operation information and the user information; the system further includes: a data analysis subunit, which is used to input the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer calculates the first attention parameter and the second attention parameter according to the attention function, and is used to determine the first intended product information according to the first attention parameter and the second intended product information according to the second attention parameter; the expressions of the first attention parameter and the second attention parameter are respectively:
[0013]
[0014]
[0015] Among them, atten1 and atten2 are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Q1 is the product browsing information, Q2 is the product operation information, K is the user information, d k is the dimension of the text vector, V1 and V2 are preset thresholds.
[0016] In some embodiments, the system stores a preset recommendation method database, which includes preset recommendation methods and preset personality types corresponding to the preset recommendation methods; the system also includes: a personality extraction unit, which is used to extract the user's personality type and the user's common language type based on the user information; a recommendation confirmation unit, which is used to confirm the preset personality type corresponding to the user's personality type in a preset communication type database based on the user's personality type, and use the preset recommendation information corresponding to the preset personality type in the recommendation method database as the recommendation method information.
[0017] In some embodiments, the intended product information includes the type of intended product and the price range of the intended product; the system also includes: a recommendation generation unit, which is used to determine a corresponding product list of intended products based on the type and price range of the intended product, and to complete the recommendation of the product list to the user terminal based on the recommendation method information.
[0018] In some embodiments, the system also includes: an optimization generation module, used to obtain transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed according to the transaction information that the user terminal has not completed the transaction, optimizing the recommendation generation module according to the feedback information to update the behavior prediction information generated by the recommendation generation module.
[0019] In a second aspect, the present application provides an e-commerce user behavior analysis method based on big data, which is applied to a server; the method comprises:
[0020] Collecting user information corresponding to the user terminal and behavior information corresponding to the user information;
[0021] Inputting the user information and the behavior information into a pre-trained behavior analysis model, wherein the behavior analysis model is used to generate behavior prediction information based on the user information and the behavior information;
[0022] The intended product information corresponding to the behavior prediction information is obtained, and recommendation method information is determined according to the user information, so as to complete the recommendation to the user terminal according to the recommendation method information.
[0023] In the third aspect, the present application provides an e-commerce user behavior analysis device based on big data, the device comprising: a user terminal and a server, the server being used to collect user information corresponding to the user terminal and behavior information corresponding to the user information to make recommendations to the user terminal; wherein the server comprises a memory and a processor: the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the e-commerce user behavior analysis method provided in any embodiment of the present application.
[0024] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the processor implements the steps of the e-commerce user behavior analysis method provided in any embodiment of the present application.
[0025] The present application provides an e-commerce user behavior analysis system based on big data, the system includes an information collection module, a data analysis module and a recommendation generation module, the user information corresponding to the user terminal collected by the information collection module and the behavior information corresponding to the user information are input into the pre-trained behavior analysis model through the data analysis module, and the behavior prediction information of the user terminal can be generated according to the user information and the behavior information corresponding to the user information, and then the recommendation generation module can obtain the intended product information corresponding to the behavior prediction information to determine the recommendation method information corresponding to the intended product information, and can recommend to the user according to the determined recommendation method information. The system provided can realize accurate prediction of user behavior in the e-commerce platform while greatly reducing computing resources, thereby improving the user transaction rate.
[0026] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0028] Figure 1 It is a structural diagram of an e-commerce user behavior analysis device based on big data provided in an embodiment of the present application;
[0029] Figure 2 This is a schematic flow chart of the steps of a method for analyzing e-commerce user behavior based on big data provided by an embodiment of the present application;
[0030] Figure 3 This is a schematic flow chart of the steps of another method for analyzing e-commerce user behavior based on big data provided by an embodiment of the present application;
[0031] Figure 4 It is a schematic block diagram of an e-commerce user behavior analysis system based on big data provided by an embodiment of the present application;
[0032] Figure 5is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0033] Figure 6 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0034] Figure 7 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0035] Figure 8 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0036] Fig. 9 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0037] Fig.10 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided in an embodiment of the present application;
[0038] Fig.11 It is a schematic block diagram of a server provided in an embodiment of the present application.
[0039] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0041] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.
[0042] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0043] It should be understood that in order to facilitate the clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, the words "first", "second" and the like are used to distinguish between the same items or similar items with substantially the same functions and effects. For example, the first fully-connected module and the second fully-connected module are only used to distinguish between different fully-connected modules, and their order is not limited. Those skilled in the art can understand that the words "first", "second" and the like do not limit the quantity and execution order, and the words "first", "second" and the like do not necessarily limit them to be different.
[0044] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0045] To facilitate understanding of the embodiments of the present application, some words involved in the embodiments of the present application are briefly explained below.
[0046] Self-attention mechanism: As the name suggests, the self-attention mechanism calculates attention for itself. That is, for each input word vector, we need to construct a self-attention word vector input, that is, the initial input of the entire model is in the form of a word vector.
[0047] For example, first multiply the word vector by three matrices to get three new vectors. The reason for multiplying the three matrix parameters instead of directly using the original word vector is to add more parameters and improve the model effect. After the input is multiplied by the three matrices, Q, K and V are obtained respectively. Q (Query) is the query requirement, K (Key) is the title of the candidate result, and V (Value) is the specific content of the candidate result. The Q of each word is matched with the K of all other words to see the relevance to each word. This relevance is weighted with the V of each word, and information is extracted from each word as needed.
[0048] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.
[0049] Currently, in e-commerce platforms, due to the limitation of the salesperson’s personal energy, a salesperson cannot receive multiple users at the same time. When multiple users consult the salesperson at the same time, it is very easy for the salesperson’s recommendation to go wrong, leading to customer loss.
[0050] In order to improve the sales efficiency of e-commerce platforms, some user behavior analysis systems are currently being applied. Data mining, machine learning and other technologies are mainly used to analyze user behavior data to discover user preferences, purchase intentions and other information, and provide corresponding optimization suggestions.
[0051] However, due to the diversity and complexity of user behavior, existing behavior analysis systems not only need to occupy too much computing resources to mine user behavior information from huge user and product data, but the accuracy of the predicted results also deviates greatly from user behavior and needs, resulting in less than ideal results for the use of behavior analysis systems on e-commerce platforms.
[0052] To solve the above problems, please refer to Figure 1 , Figure 1 Schematic diagram of the structure of an e-commerce user behavior analysis device based on big data provided by an embodiment of the present application. Figure 1 As shown, the provided e-commerce user behavior analysis device 10 includes a user terminal 11 and a server 12 , and the server 12 is used to collect user information corresponding to the user terminal 11 and behavior information corresponding to the user information to complete recommendations to the user terminal 11 .
[0053] Specifically, in the e-commerce user behavior analysis device 10 provided in the present application, when a user logs in to the server 12 corresponding to the e-commerce system through the user terminal 11, the server 12 can obtain the user information corresponding to the user terminal 11 and the behavior information corresponding to the user information under the conditions permitted by relevant laws and regulations after the user terminal 11 authorizes. Then the server 12 can analyze the user information and behavior information to obtain the behavior prediction information corresponding to the user terminal. The server 12 then determines the recommendation method information based on the intended product information corresponding to the behavior prediction information and the user information, so that the server 12 can recommend to the user terminal in the most appropriate recommendation method to improve the transaction rate of the user terminal.
[0054] In some embodiments, the user terminal 11 may be a computer, a mobile phone, or a tablet computer. As long as the user can log in to any terminal device of the server 12 corresponding to the e-commerce system, it can be used as the user terminal 11 provided in the embodiment of the present application, so the present application does not limit the type of the user terminal 11.
[0055] The present application provides an e-commerce user behavior analysis device based on big data, which can accurately predict the behavior of user terminals in an e-commerce platform while significantly reducing computing resources, thereby improving the transaction rate of user terminals.
[0056] Please refer to Figure 2 , Figure 2The method is applied to the server of the e-commerce user behavior analysis device based on big data provided in any embodiment of the present application.
[0057] like Figure 2 As shown, the provided method includes steps S101 to S103.
[0058] S101. Collect user information corresponding to the user terminal and behavior information corresponding to the user information.
[0059] Specifically, when the user terminal logs in to the e-commerce platform corresponding to the server, the server can obtain the user information corresponding to the user terminal and the behavior information corresponding to the user information, such as obtaining the user's registration information and the user's operation record information on the e-commerce platform. The server can then accurately predict the user's behavior based on the user's operation on the e-commerce platform and the user information, thereby improving the transaction rate of the user terminal on the e-commerce platform.
[0060] S102. Input the user information and behavior information into a pre-trained behavior analysis model, and the behavior analysis model is used to generate behavior prediction information based on the user information and behavior information.
[0061] Specifically, the server inputs the user information and the behavior information corresponding to the user information into the behavior analysis model pre-trained by the server, such as the multi-layer neural network model pre-trained by the server. Then the server can obtain the behavior analysis model to generate behavior prediction information according to the user information and behavior information. Therefore, using the provided method, the server can combine the user's operation with the user's personal information through the behavior analysis model, and more comprehensively measure the deep meaning of the user's behavior on the e-commerce platform. Finally, the server can accurately predict the user's behavior prediction information on the e-commerce platform, so that the server can better recommend it according to the user's purpose.
[0062] S103. Obtain intended product information corresponding to the behavior prediction information, determine recommendation method information according to the user information, and complete the recommendation to the user terminal according to the recommendation method information.
[0063] Specifically, after obtaining the user's behavior prediction information, the server can confirm the user's intended product information on the e-commerce platform based on the behavior prediction information. At the same time, in order to improve user satisfaction, different users should correspond to different recommendation methods, so the server needs to confirm the corresponding recommendation method information based on the user information. The recommendation method information may include the recommendation method for the server to recommend the intended product information to the user terminal, such as voice push or text push, and may also include the recommendation style for recommending product information to the user terminal, such as focusing on product parameter introduction or emotion-driven recommendation. The server can then generate personalized recommendations for the intended products for each user terminal, thereby improving the transaction rate of the user terminal.
[0064] In some embodiments, please refer to Figure 3 , Figure 3 1 is a flowchart showing the steps of another method for analyzing e-commerce user behavior based on big data provided by an embodiment of the present application. Figure 2 The embodiment provided is different in that after step S101 and before step S102, Figure 3 The provided embodiment also includes steps S104 and S105.
[0065] S104. Confirm missing data information, blank data information, and white noise data information corresponding to the user information and the behavior information according to the user information and the behavior information.
[0066] S105. Fill in the missing data information, delete the blank data information, and denoise the white noise data information according to the data type corresponding to the missing data information, so as to complete the data cleaning of the user information and behavior information.
[0067] Since the user information and behavior information collected by the server are not necessarily complete data, the use of incomplete data for behavior prediction will result in a decrease in the accuracy of behavior prediction. Therefore, the server determines whether there is missing data information, blank data information, and white noise data information in the collected data, for example, by detecting the received information according to the preset format of user information and behavior information, and then determining whether there is missing, blank, and white noise.
[0068] After the server confirms the missing, blank and white noise data in the user information and behavior information, the missing data information is supplemented by a data cleaning algorithm, such as a binning method or a regression method, and the blank data information is directly deleted. For white noise data, the white noise in the data is removed by methods such as the Kalman filter algorithm, the Gaussian filter algorithm and the low-pass filter algorithm.
[0069] In some embodiments, the behavior analysis model includes a preset multi-layer neural network, which includes at least a fully connected layer and a coding layer; the behavior information includes product browsing information and product operation information of the user terminal; the user information and behavior information are input into the pre-trained behavior analysis model, and the behavior analysis model is used to generate behavior prediction information based on the user information and behavior information, including: inputting the user information and the product browsing information and product operation information of the user terminal into the fully connected layer for full connection, inputting the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer outputs the behavior prediction information based on the self-attention mechanism.
[0070] Since both user information and behavior information are text information or list information, the user information and behavior information are fully connected by using a fully connected layer, and then the fully connected user information and behavior information are input into the encoding layer. Then the encoding layer can mine the correlation between user information, product browsing information, product operation information and behavior prediction information based on the self-attention mechanism, so that the server can accurately predict the user's behavior. The provided method conducts self-supervisory training on user information, product browsing information and product operation information by building an intention product generation model, which can fully mine the risk co-occurrence relationship and latent space logical association relationship between the two sets of data, namely user information and product browsing information, and user information and product operation information.
[0071] Exemplarily, the coding layer stores an attention function, and the attention function is used to calculate a first attention parameter of product browsing information and user information, and a second attention parameter of product operation information and user information; the fully connected user information, product browsing information, and product operation information are input into the coding layer, and the coding layer outputs behavior prediction information based on a self-attention mechanism, including: the fully connected user information, product browsing information, and product operation information are input into the coding layer, and the coding layer calculates a first attention parameter and a second attention parameter according to the attention function, and is used to determine the first behavior prediction information according to the first attention parameter and the second behavior prediction information according to the second attention parameter;
[0072] The expressions of the first attention parameter and the second attention parameter are:
[0073]
[0074]
[0075] Among them, atten1 and atten2 are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Q1 is the product browsing information, Q2 is the product operation information, K is the user information, d k is the dimension of the text vector, V1 and V2 are preset thresholds.
[0076] When the preset multi-layer neural network provided in this application is a Transformer model, Transformer is a model that uses an attention mechanism to increase the model training speed. Trasnformer can be said to be a deep learning model based entirely on the self-attention mechanism, because it is suitable for parallel computing, and the complexity of its own model causes it to be higher in accuracy and performance than the previously popular RNN recurrent neural network. The provided method constructs a first attention function and a second attention function by means of Q, K, and V vectors to calculate multiple groups of first attention parameters for calculating product browsing information and user information, and second attention parameters for product operation information and user information.
[0077] By constructing the Softmax function to obtain the value of the attention function, Q, K, and V construct multi-head attention through multiple sets of weights. Then the calculation of the first association parameter, the second association parameter, and the third association parameter is completed to obtain the first behavior prediction information corresponding to the first association parameter and the second behavior prediction information corresponding to the second association parameter. Then the behavior prediction information corresponding to the user terminal can be obtained in three dimensions to make more comprehensive product recommendations to users.
[0078] In some embodiments, the server stores a preset recommendation method database, which includes preset recommendation methods and preset personality types corresponding to the preset recommendation methods; determining the recommendation method information based on user information includes: extracting the user's personality type and the user's common language type based on the user information; confirming the preset personality type corresponding to the user's personality type in a preset communication type database based on the user's personality type, obtaining the preset recommendation information corresponding to the preset personality type in the recommendation method database, and generating the recommendation method information based on the intended product information and the preset recommendation information.
[0079] Due to the different personalities of different users, many of their behavioral preferences in daily life vary greatly. Therefore, if recommendations are not made to users based on their personalities, it is easy to lead to poor user experience and ultimately failure to close a deal. Therefore, the provided method extracts the user's personality type and the user's common language type based on user information, such as introversion and extroversion, or extracts features that can be used to analyze the user's personality type, such as constellations, age, MBTI classification (Myers Briggs Type Indicator) and other information. The user's common language type can be different dialects such as Mandarin and Cantonese, or it can be language information from different countries such as Chinese and English. In the provided method, the server can confirm the preset personality type corresponding to the user's personality type in the preset communication type database according to the user's personality type, and obtain the preset recommendation information corresponding to the preset personality type in the recommendation method database. Finally, the server generates recommendation method information based on the intended product information and the preset recommendation information, completes the targeted recommendation for the user, and improves the user's user experience and the final transaction rate.
[0080] In some embodiments, the intended product information includes the type of intended product and the price range of the intended product; the acquisition of the intended product information corresponding to the behavior prediction information includes: determining the corresponding product list of intended products based on the type and price range of the intended product, which is used to complete the recommendation of the product list to the user terminal based on the recommendation method information.
[0081] The server can confirm the type of the user's intended goods and the price range of the intended goods through user behavior information, such as the price range selected by the user terminal in the e-commerce platform or the average price range of the goods that the user clicks to browse. Then, the product list of the user's intended goods on the e-commerce platform can be determined according to the type and price range of the intended goods. For example, if the user's intended product type is (autumn clothing, women's clothing, warm colors, classical style), and the price range is 200 to 400 yuan, then the server can quickly determine the product list corresponding to the user terminal according to the user's product type and price range, so that the user can have more choices and improve the user's experience.
[0082] In some embodiments, after obtaining the intended product information corresponding to the behavior prediction information and determining the recommendation method information based on the user information, and completing the recommendation to the user terminal based on the recommendation method information, the method also includes: obtaining the transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed according to the transaction information that the user terminal has not completed the transaction, optimizing the recommendation generation module according to the feedback information to update the behavior prediction information generated by the recommendation generation module.
[0083] The provided method server can timely update the recommendation generation module according to the user's feedback each time after recommending the intended product to the user. This allows the recommendation generation module to more accurately predict the behavior of the user corresponding to the user terminal. At the same time, the server can also generate modification information of the e-commerce platform according to the user's feedback information, and the modification information is used to modify the product pricing, title and other information to increase the attractiveness of the products in the e-commerce platform to users.
[0084] The present application provides an e-commerce user behavior analysis method based on big data, which can achieve accurate prediction of user behavior in the e-commerce platform while significantly reducing computing resources, thereby improving user transaction rates.
[0085] Please refer to Figure 4 , Figure 4 This is a schematic block diagram of an e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. The system is used to execute the steps of the aforementioned e-commerce user behavior analysis method based on big data.
[0086] like Figure 4 As shown, the provided e-commerce user behavior analysis system 200 includes an information collection module 201 , a data analysis module 202 and a recommendation generation module 203 .
[0087] The information collection module 201 is used to collect user information corresponding to the user terminal and behavior information corresponding to the user information.
[0088] The data analysis module 202 is used to input the user information and the behavior information into a pre-trained behavior analysis model, and the behavior analysis model is used to generate behavior prediction information based on the user information and the behavior information.
[0089] The recommendation generation module 203 is used to obtain the intended product information corresponding to the behavior prediction information, determine the recommendation method information according to the user information, and complete the recommendation to the user terminal according to the recommendation method information.
[0090] In some embodiments, please refer to Figure 5 , Figure 5 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 4 The embodiment provided is different in that Figure 5 The provided user behavior analysis system 200 also includes an information confirmation module 204 and a data cleaning module 205 .
[0091] The information confirmation module 204 is used to confirm the missing data information, blank data information and white noise data information corresponding to the user information and the behavior information according to the user information and the behavior information.
[0092] The data cleaning module 205 is used to fill the missing data information, delete the blank data information, and denoise the white noise data information according to the data type corresponding to the missing data information, so as to complete the data cleaning of the user information and the behavior information.
[0093] In some embodiments, the behavior analysis model includes a preset multi-layer neural network, and the preset multi-layer neural network includes at least a fully connected layer and an encoding layer; please refer to Figure 6 , Figure 6 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 4 The embodiment provided is different in that Figure 6 In the provided user behavior analysis system 200, the information collection module 201 includes an information collection unit 201a, and the data analysis module 202 includes a data analysis unit 202a.
[0094] The information collection unit 201a is used to collect the user information and the commodity browsing information and commodity operation information of the user terminal.
[0095] The data analysis unit 202a is used to input the user information and the product browsing information and product operation information of the user terminal into the fully connected layer for full connection, and input the fully connected user information, product browsing information and product operation information into the encoding layer, and the encoding layer outputs the behavior prediction information based on the self-attention mechanism.
[0096] Exemplarily, the coding layer stores an attention function, and the attention function is used to calculate a first attention parameter of the product browsing information and the user information, and a second attention parameter of the product operation information and the user information. Figure 7 , Figure 7 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 6 The embodiment provided is different in that Figure 7 In the provided user behavior analysis system 200, the data analysis unit 202a includes a data analysis sub-unit 202a1.
[0097] The data analysis subunit 202a1 is used to input the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer calculates the first attention parameter and the second attention parameter according to the attention function, and is used to determine the first intended product information according to the first attention parameter and the second intended product information according to the second attention parameter; the expressions of the first attention parameter and the second attention parameter are respectively:
[0098]
[0099]
[0100] Among them, atten1 and atten2 are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Q1 is the product browsing information, Q2 is the product operation information, K is the user information, d k is the dimension of the text vector, V1 and V2 are preset thresholds.
[0101] In some embodiments, the e-commerce user behavior analysis system 200 stores a preset recommendation method database, which includes preset recommendation methods and preset personality types corresponding to the preset recommendation methods; please refer to Figure 8 , Figure 8 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 4 The embodiment provided is different in that Figure 8 In the provided user behavior analysis system 200, the recommendation generation module 203 further includes a personality extraction unit 203a and a recommendation confirmation unit 203b.
[0102] The personality extraction unit 203a is used to extract the personality type of the user and the type of common language used by the user according to the user information.
[0103] The recommendation confirmation unit 203b is used to confirm the preset personality type corresponding to the user personality type in the preset communication type database according to the user personality type, and use the preset recommendation information corresponding to the preset personality type in the recommendation method database as the recommendation method information.
[0104] In some embodiments, please refer to Fig. 9 , Fig. 9 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 4 The embodiment provided is different in that Figure 8 In the provided user behavior analysis system 200, the recommendation generation module 203 further includes a recommendation generation unit 203c.
[0105] The recommendation generating unit 203c is used to determine a product list of the intended products according to the type and price range of the intended products, and to recommend the product list to the user terminal according to the recommendation method information.
[0106] In some embodiments, please refer to Fig.10 , Fig.10 is a schematic block diagram of another e-commerce user behavior analysis system based on big data provided by an embodiment of the present application. Figure 4 The embodiment provided is different in that Fig.10 The provided user behavior analysis system 200 also includes an optimization generation module 206 .
[0107] The optimization generation module 206 is used to obtain the transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed that the user terminal has not completed the transaction based on the transaction information, the recommendation generation module is optimized based on the feedback information to update the behavior prediction information generated by the recommendation generation module.
[0108] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.
[0109] The above-mentioned device can be implemented in the form of a computer program. Fig.11 Run on the server shown.
[0110] This application provides a server. Fig.11 As shown, Fig.11 1 is a schematic block diagram of a server provided in an embodiment of the present application. The server may include a processor, a memory, and a network interface. The processor, the memory, and the network interface are connected via a system bus, such as an I2C (Inter-integrated Circuit) bus.
[0111] Specifically, the processor may be a micro-controller unit (MCU), a central processing unit (CPU) or a digital signal processor (DSP).
[0112] Specifically, the memory may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.
[0113] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Fig.11 The structure shown in the figure is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the server to which the solution of the present application is applied. The specific server may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.
[0114] The processor is used to run a computer program stored in the memory, and implement corresponding steps in any one of the big data-based e-commerce user behavior analysis methods provided in the embodiments of the present application when executing the computer program.
[0115] The exemplary processor provided is used to implement the following steps:
[0116] Collect user information corresponding to the user terminal and behavior information corresponding to the user information.
[0117] The user information and the behavior information are input into a pre-trained behavior analysis model, and the behavior analysis model is used to generate behavior prediction information based on the user information and the behavior information.
[0118] The intended product information corresponding to the behavior prediction information is obtained, and recommendation method information is determined according to the user information, so as to complete the recommendation to the user terminal according to the recommendation method information.
[0119] In some embodiments, before the user information and behavior information are input into the pre-trained behavior analysis model, the provided processor is also used to implement: confirming the missing data information, blank data information and white noise data information corresponding to the user information and behavior information based on the user information and behavior information; filling the missing data information, deleting the blank data information, and denoising the white noise data information according to the data type corresponding to the missing data information, thereby completing the data cleaning of the user information and behavior information.
[0120] In some embodiments, the behavior analysis model includes a preset multi-layer neural network, which includes at least a fully connected layer and a coding layer; the behavior information includes product browsing information and product operation information of the user terminal; the provided processor is used to implement: inputting the user information and the product browsing information and product operation information of the user terminal into the fully connected layer for full connection, inputting the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer outputs behavior prediction information based on the self-attention mechanism.
[0121] In some embodiments, the coding layer stores an attention function, and the attention function is used to calculate a first attention parameter of product browsing information and user information, and a second attention parameter of product operation information and user information; the provided processor is used to implement: inputting the fully connected user information, product browsing information and product operation information into the coding layer, and the coding layer calculates the first attention parameter and the second attention parameter according to the attention function, and is used to determine the first behavior prediction information according to the first attention parameter and the second behavior prediction information according to the second attention parameter; the expressions of the first attention parameter and the second attention parameter are respectively:
[0122]
[0123]
[0124] Among them, atten1 and atten2 are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Q1 is the product browsing information, Q2 is the product operation information, K is the user information, d k is the dimension of the text vector, V1 and V2 are preset thresholds.
[0125] In some embodiments, the server stores a preset recommendation method database, which includes preset recommendation methods and preset personality types corresponding to the preset recommendation methods; the provided processor is used to implement: extracting the user's personality type and the user's common language type based on user information; confirming the preset personality type corresponding to the user's personality type in a preset communication type database based on the user's personality type, obtaining the preset recommendation information corresponding to the preset personality type in the recommendation method database, and generating recommendation method information based on the intended product information and the preset recommendation information.
[0126] In some embodiments, the intended product information includes the type of the intended product and the price range of the intended product; the provided processor is used to implement: determining the corresponding product list of the intended product according to the type and price range of the intended product, and completing the recommendation of the product list to the user terminal according to the recommendation method information.
[0127] In some embodiments, after obtaining the intended product information corresponding to the behavior prediction information and determining the recommendation method information based on the user information, and completing the recommendation to the user terminal based on the recommendation method information, the provided processor is also used to implement: obtaining the transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed according to the transaction information that the user terminal has not completed the transaction, optimizing the recommendation generation module according to the feedback information to update the behavior prediction information generated by the recommendation generation module.
[0128] It should be noted that technical personnel in the relevant field can clearly understand that for the convenience and brevity of description, the specific working process of the server described above can refer to the corresponding process in the aforementioned embodiment of the e-commerce user behavior analysis method based on big data, and will not be repeated here.
[0129] A computer-readable storage medium is also provided in an embodiment of the present application, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and the processor executes the program instructions to implement the steps of the big data-based e-commerce user behavior analysis method provided in the above embodiment.
[0130] The computer-readable storage medium may be an internal storage unit of the server in the aforementioned embodiment, such as a hard disk or memory of the server. The computer-readable storage medium may also be an external storage device of the server, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., provided on the server.
[0131] Since the computer program stored in the computer-readable storage medium can execute any one of the e-commerce user behavior analysis methods based on big data provided in the embodiments of the present application, the beneficial effects that can be achieved by any one of the e-commerce user behavior analysis methods based on big data provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.
[0132] The serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments. The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.
Claims
1. An e-commerce user behavior analysis system based on big data, characterized in that: The system comprises: An information collection module, used to collect user information corresponding to the user terminal and behavior information corresponding to the user information; the user information includes registration information, and the behavior information includes operation record information; A data analysis module, used for inputting the user information and the behavior information into a pre-trained behavior analysis model, wherein the behavior analysis model is used for generating behavior prediction information according to the user information and the behavior information; a recommendation generation module, used to obtain the intended product information corresponding to the behavior prediction information, determine the recommendation method information according to the user information, and complete the recommendation to the user terminal according to the recommendation method information; the recommendation method information includes the recommendation method and the recommendation style, the recommendation method includes voice push or text push, and the recommendation style includes product parameter introduction recommendation or emotion-driven recommendation; The behavior analysis model includes a preset multi-layer neural network, and the preset multi-layer neural network includes at least a fully connected layer and a coding layer; the system also includes: An information collection unit, used to collect the user information and the commodity browsing information and commodity operation information of the user terminal; A data analysis unit, configured to input the user information and the commodity browsing information and commodity operation information of the user terminal into the fully connected layer for full connection, and input the fully connected user information, commodity browsing information and commodity operation information into the encoding layer, and the encoding layer outputs the behavior prediction information based on a self-attention mechanism; The coding layer stores an attention function, which is used to calculate a first attention parameter of the product browsing information and the user information, and a second attention parameter of the product operation information and the user information; the system further includes: a data analysis subunit, configured to input the fully connected user information, product browsing information, and product operation information into the coding layer, wherein the coding layer calculates the first attention parameter and the second attention parameter according to the attention function, and determines the first behavior prediction information according to the first attention parameter and the second behavior prediction information according to the second attention parameter; The expressions of the first attention parameter and the second attention parameter are respectively: in, and are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Browse information for products, is the product operation information, K is the user information, is the dimension of the text vector, and is the preset threshold; The system also includes: an optimization generation module, which is used to obtain transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed according to the transaction information that the user terminal has not completed the transaction, optimizing the recommendation generation module according to the feedback information to update the behavior prediction information generated by the recommendation generation module; generating modification information according to the feedback information, and the modification information is used to modify the product pricing and title.
2. The system according to claim 1, characterized in that The system further comprises: An information confirmation module, used to confirm missing data information, blank data information and white noise data information corresponding to the user information and the behavior information according to the user information and the behavior information; A data cleaning module is used to fill in the missing data information, delete the blank data information, and denoise the white noise data information according to the data type corresponding to the missing data information, so as to complete the data cleaning of the user information and the behavior information.
3. The system according to claim 1, characterized in that The system stores a preset recommendation method database, which includes preset recommendation methods and preset personality types corresponding to the preset recommendation methods; the system also includes: A personality extraction unit, used to extract the user's personality type and the user's common language type according to the user information; A recommendation confirmation unit is used to confirm the preset personality type corresponding to the user personality type in a preset communication type database according to the user personality type, obtain the preset recommendation information corresponding to the preset personality type in the recommendation method database, and generate the recommendation method information according to the intended product information and the preset recommendation information.
4. The system according to claim 1, characterized in that The system further comprises: The optimization generation module is used to obtain the transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed that the user terminal has not completed the transaction based on the transaction information, the recommendation generation module is optimized based on the feedback information to update the behavior prediction information generated by the recommendation generation module.
5. An e-commerce user behavior analysis method based on big data, characterized in that: Applied to a server; the method comprises: Collecting user information corresponding to the user terminal and behavior information corresponding to the user information; the user information includes registration information, and the behavior information includes operation record information; Inputting the user information and the behavior information into a pre-trained behavior analysis model, wherein the behavior analysis model is used to generate behavior prediction information based on the user information and the behavior information; Acquire the intended product information corresponding to the behavior prediction information, determine the recommendation method information according to the user information, and complete the recommendation to the user terminal according to the recommendation method information; the recommendation method information includes the recommendation method and the recommendation style, the recommendation method includes voice push or text push, and the recommendation style includes product parameter introduction-oriented recommendation or emotion-driven recommendation; The behavior analysis model includes a preset multi-layer neural network, and the preset multi-layer neural network includes at least a fully connected layer and a coding layer; the method also includes: Collecting the user information and the commodity browsing information and commodity operation information of the user terminal; Inputting the user information and the commodity browsing information and commodity operation information of the user terminal into the fully connected layer for full connection, inputting the fully connected user information, commodity browsing information and commodity operation information into the encoding layer, and the encoding layer outputs the behavior prediction information based on the self-attention mechanism; The coding layer stores an attention function, and the attention function is used to calculate a first attention parameter of the product browsing information and the user information, and a second attention parameter of the product operation information and the user information; the method further includes: The fully connected user information, product browsing information, and product operation information are input into the encoding layer, and the encoding layer calculates the first attention parameter and the second attention parameter according to the attention function, so as to determine the first behavior prediction information according to the first attention parameter and the second behavior prediction information according to the second attention parameter; The expressions of the first attention parameter and the second attention parameter are respectively: in, and are the first attention parameter and the second attention parameter respectively, Softmax is the normalization function, Browse information for products, is the product operation information, K is the user information, is the dimension of the text vector, and is the preset threshold; The method also includes: obtaining transaction information and feedback information of the user terminal after completing the recommendation; if it is confirmed according to the transaction information that the user terminal has not completed the transaction, optimizing the recommendation generation module according to the feedback information to update the behavior prediction information generated by the recommendation generation module; generating modification information according to the feedback information, and the modification information is used to modify the product pricing and title.
6. An e-commerce user behavior analysis device based on big data, characterized in that: The device comprises: A user terminal and a server, the server being used to collect user information corresponding to the user terminal and behavior information corresponding to the user information to complete the recommendation to the user terminal; Wherein, the server includes a memory and a processor: the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the e-commerce user behavior analysis method as described in claim 5.
7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the processor implements the steps of the e-commerce user behavior analysis method as claimed in claim 5.
