E-commerce online operation data analysis method and device and computer equipment

Abstract

The invention provides an e-commerce online operation data analysis method and device and computer equipment. The method comprises the steps of obtaining operation data of an e-commerce platform; wherein the operation data comprises a commodity browsing record, a user transaction record and a user behavior log; analyzing the commodity browsing record and the user transaction record based on a preset user portrait analysis model to obtain a user initial portrait; matching corresponding model difference parameters in a database based on the initial portrait of the user; obtaining an initial data analysis model, and performing differentiation adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model; and analyzing the user behavior log based on the adjusted data analysis model to obtain a corresponding analysis result. According to the method and the device, more efficient operation decision and personalized user experience improvement are realized, and the efficiency and the accuracy of data analysis are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an analysis method, device and computer equipment for online operation data of e-commerce. Background Art

[0002] The analysis of online operation data has become a key link for e-commerce platforms to enhance their competitiveness, optimize user experience, and achieve precision marketing. With the surge in Internet users and the diversification of consumer behavior, e-commerce platforms generate massive amounts of data every day, covering multiple dimensions such as users' browsing habits, purchasing behaviors, and changes in preferences. Effective use of the above data can not only provide insights into market trends, but also provide a deep understanding of user needs, providing strong data support for strategies such as personalized recommendations, inventory management, and advertising.

[0003] However, traditional data analysis methods often face several challenges: first, the amount of data is huge and complex, and how to efficiently process and analyze the above data has become a major problem; second, user behavior is changeable, and simple statistical analysis is difficult to capture the personalized characteristics of users, resulting in low accuracy of the recommendation system; third, general data analysis models may not be able to fully consider the subtle differences between different user groups, affecting the accuracy and practicality of the analysis results; therefore, current analysis methods usually have the defect of low accuracy. Summary of the invention

[0004] The main purpose of the present invention is to provide an analysis method, device and computer equipment for online operation data of e-commerce, aiming to overcome the defect of low accuracy of current analysis methods.

[0005] To achieve the above object, the present invention provides an analysis method for online e-commerce operation data, comprising the following steps:

[0006] Obtaining the operation data of the e-commerce platform; wherein the operation data includes product browsing records, user transaction records, and user behavior logs;

[0007] Analyze the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait;

[0008] Based on the initial user portrait, the corresponding model difference parameters are matched in the database; wherein the database pre-stores a mapping relationship between the user portrait and the model difference parameters; the model difference parameters include differentiated model parameters and their corresponding difference values;

[0009] Acquire an initial data analysis model, and perform differential adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model;

[0010] The user behavior log is analyzed based on the adjusted data analysis model to obtain corresponding analysis results.

[0011] Furthermore, the step of analyzing the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait includes:

[0012] Based on the preset user portrait analysis model, feature extraction is performed on the product browsing record and the user transaction record to obtain browsing record features and transaction record features;

[0013] Combining the browsing record features with the transaction record features to obtain a portrait feature;

[0014] A classification operation is performed on the portrait features to obtain a corresponding classification label as the initial portrait of the user.

[0015] Furthermore, after the step of analyzing the user behavior log based on the adjusted data analysis model to obtain the corresponding analysis result, the step further includes:

[0016] The operation data and the analysis results are associated and stored in a database.

[0017] Furthermore, the step of differentially adjusting the model parameters of the initial data analysis model based on the model difference parameters includes:

[0018] Based on the model difference parameters, matching differentiated target model parameters from the initial data analysis model;

[0019] Based on the difference values ​​in the model difference parameters, various target model parameters are adjusted.

[0020] Furthermore, the step of obtaining the initial data analysis model includes:

[0021] Obtain the platform name information of the e-commerce platform and the latest update time of the platform;

[0022] Generate a request code based on the platform name information and the latest update time;

[0023] The request code is sent to the management end, and an initial data analysis model matching the request code is matched in the management end; wherein the management end pre-stores a mapping relationship between a plurality of different request codes and data analysis models.

[0024] Furthermore, the step of generating a request code based on the platform name information and the latest update time includes:

[0025] Mapping each Chinese character in the platform name information into a corresponding coded character; wherein the platform name information is a Chinese name;

[0026] Create a blank data table with multiple rows and columns;

[0027] Add the coded characters mapped to each Chinese character to the blank data table one by one in order to obtain a character data table; wherein only one character is added to a table;

[0028] Based on the year in the latest update time, the character data table is replaced to obtain a transformation data table; based on the month and date in the latest update time, a simulation curve is generated;

[0029] The simulation curve is superimposed on the transformation data table according to a preset rule, all characters in the transformation data table located below the simulation curve are obtained, and they are combined in sequence, and the obtained character combination is used as the request code.

[0030] Furthermore, the step of replacing the character data table based on the year in the latest update time to obtain a transformation data table; and generating a simulation curve based on the month and date in the latest update time includes:

[0031] Obtain the year in the latest update time, and add the characters in the year in sequence to a data table with two rows and two columns to obtain a year data table; wherein the year is a four-digit number;

[0032] Superimposing the center of the year data table and the character data table, and superimposing the characters in the year data table to the overlapping positions in the character data table to obtain a transformation data table;

[0033] Obtain the month and date in the latest update time, create a coordinate system, use the number corresponding to the month as the horizontal coordinate of the feature point, and use the number corresponding to the date as the vertical coordinate of the feature point;

[0034] Based on the feature points and the origin of the coordinate system, a straight line is constructed as the simulation curve.

[0035] Furthermore, the step of mapping each Chinese character in the platform name information into a corresponding coded character includes:

[0036] Obtain a standard coding table;

[0037] Perform keyword recognition on the platform name information to obtain a key name;

[0038] Rearranging the standard coding table based on the key name to obtain a rearranged coding table;

[0039] Based on the rearranged encoding table, each Chinese character in the platform name information is mapped to a corresponding encoding character.

[0040] The present invention also provides an analysis device for online e-commerce operation data, comprising:

[0041] An acquisition unit, used to acquire operation data of the e-commerce platform; wherein the operation data includes commodity browsing records, user transaction records, and user behavior logs;

[0042] A first analysis unit, configured to analyze the commodity browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait;

[0043] A matching unit, configured to match corresponding model difference parameters in a database based on the initial user portrait; wherein a mapping relationship between the user portrait and the model difference parameters is pre-stored in the database; and the model difference parameters include differentiated model parameters and their corresponding difference values;

[0044] An adjustment unit, configured to obtain an initial data analysis model, and perform differential adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model;

[0045] The second analysis unit is used to analyze the user behavior log based on the adjusted data analysis model to obtain corresponding analysis results.

[0046] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0047] The analysis method, device and computer equipment of online operation data of e-commerce provided by the present invention include: obtaining the operation data of the e-commerce platform; wherein the operation data includes commodity browsing records, user transaction records, and user behavior logs; analyzing the commodity browsing records and user transaction records based on a preset user portrait analysis model to obtain the user initial portrait; matching the corresponding model difference parameters in the database based on the user initial portrait; wherein the database pre-stores a mapping relationship between the user portrait and the model difference parameter; the model difference parameter includes a differentiated model parameter and its corresponding difference value; obtaining the initial data analysis model, and differentially adjusting the model parameters of the initial data analysis model based on the model difference parameter to obtain the adjusted data analysis model; analyzing the user behavior log based on the adjusted data analysis model to obtain the corresponding analysis result. In the present invention, by integrating multi-source data such as user browsing records, transaction records and behavior logs, a personalized analysis model is constructed using advanced user portrait technology. The method uses a preset user portrait analysis model to deeply mine the original data and form an initial user portrait. On this basis, the concept of model difference parameters was further introduced. According to the slight differences in user characteristics, the most matching differentiated adjustment parameters were found in the database, and the basic analysis model was customized and optimized to ensure that the analysis model could better fit the characteristics of different user groups. Finally, the adjusted model was used to conduct in-depth analysis of user behavior logs to obtain more accurate and detailed analysis results, thereby helping e-commerce platforms to achieve more efficient operational decisions and personalized user experience, and improving the efficiency and accuracy of data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is a schematic diagram of steps of a method for analyzing online operation data of e-commerce in one embodiment of the present invention;

[0049] Figure 2 It is a structural block diagram of an analysis device for online e-commerce operation data in one embodiment of the present invention;

[0050] Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0051] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0053] Reference Figure 1 In one embodiment of the present invention, a method for analyzing online operation data of e-commerce is provided, comprising the following steps:

[0054] Step S1, obtaining the operation data of the e-commerce platform; wherein the operation data includes product browsing records, user transaction records, and user behavior logs;

[0055] Step S2, analyzing the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait;

[0056] Step S3, based on the initial user portrait, matching the corresponding model difference parameters in the database; wherein the database pre-stores a mapping relationship between the user portrait and the model difference parameters; the model difference parameters include differentiated model parameters and their corresponding difference values;

[0057] Step S4, obtaining an initial data analysis model, and differentially adjusting the model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model;

[0058] Step S5, analyzing the user behavior log based on the adjusted data analysis model to obtain corresponding analysis results.

[0059] In this embodiment, it specifically includes:

[0060] Step S1: Obtaining the operational data of the e-commerce platform

[0061] Data sources: First, data is collected from the backend system of the e-commerce platform, including but not limited to product browsing records (details, categories, time, etc. of products viewed by users), user transaction records (purchased products, quantity, payment methods, transaction time, etc.) and user behavior logs (such as search keywords, page dwell time, click-through rate, add to cart operations, etc.). The above data is usually stored in a database or big data warehouse and needs to be obtained through an API interface or data extraction tool.

[0062] Step S2: User initial profile analysis

[0063] Analysis model: Use preset user portrait analysis models, such as collaborative filtering, clustering algorithms, or deep learning models, to analyze product browsing records and user transaction records. The above models can identify user behavior patterns, preferences, and purchasing habits.

[0064] Generate a profile: Through model analysis, we can summarize the user's preliminary characteristics, such as age range, gender orientation, consumption level, interest areas, etc., to form an initial user profile. This step is the basis for subsequent personalized analysis.

[0065] Step S3: Matching model difference parameters

[0066] Database mapping: In the pre-built database, the mapping relationship between different user portrait features and corresponding model adjustment parameters is stored. The above difference parameters are designed to adapt to different user groups and can reflect the special needs or behavioral deviations of specific user groups.

[0067] Matching process: Based on the initial user profile, query the database for the corresponding optimal model difference parameters. The above parameters can be adjusted model sensitivity, weight distribution, or other hyperparameters in the algorithm to better match the specificity of the user group.

[0068] Step S4: Adjust the data analysis model

[0069] Initial Model: Select one or more initial data analysis models suitable for e-commerce analysis, such as time series analysis, association rule mining, or predictive models.

[0070] Parameter adjustment: Using the model difference parameters obtained from step S3, the parameters of the initial model are adjusted differentially. This involves adjusting the threshold of the prediction algorithm, optimizing the decision boundary of the classifier, or changing the feature selection criteria to more accurately reflect the behavior of the target user group.

[0071] Step S5: Analyze user behavior logs

[0072] Apply the adjusted model: Apply the adjusted data analysis model to the user behavior log for in-depth analysis. This step aims to discover potential patterns in user behavior, predict future trends, or identify abnormal behavior.

[0073] Analysis results: Through model analysis, a series of analysis results can be output, such as user group segmentation, purchase intention prediction, potential high-value user identification, product recommendation optimization suggestions, etc. The above information has direct guiding significance for e-commerce marketing, inventory management, user experience optimization and other aspects.

[0074] In summary, this technical solution integrates multi-source data such as user browsing records, transaction records, and behavior logs, and uses advanced user portrait technology to build a personalized analysis model. This method uses a preset user portrait analysis model to deeply mine the original data to form an initial user portrait. On this basis, the concept of model difference parameters is further introduced. According to the slight differences in user characteristics, the most matching differentiated adjustment parameters are found in the database, and the basic analysis model is customized and optimized to ensure that the analysis model can better fit the characteristics of different user groups. Finally, the adjusted model is used to conduct in-depth analysis of user behavior logs to obtain more accurate and detailed analysis results, thereby helping e-commerce platforms to achieve more efficient operational decisions and personalized user experience improvements, and improving the efficiency and accuracy of data analysis.

[0075] In one embodiment, the step of analyzing the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait includes:

[0076] Based on the preset user portrait analysis model, feature extraction is performed on the product browsing record and the user transaction record to obtain browsing record features and transaction record features;

[0077] Combining the browsing record features with the transaction record features to obtain a portrait feature;

[0078] A classification operation is performed on the portrait features to obtain a corresponding classification label as the initial portrait of the user.

[0079] In this embodiment, it specifically includes:

[0080] Product browsing history feature extraction:

[0081] Browsing frequency: Counts the frequency with which users browse each product category and reveals users’ areas of interest.

[0082] Dwell time: records the average length of time users stay on each product page, reflecting the user's attention to the product.

[0083] Browsing time distribution: Analyze users’ active time periods to understand their online behavior habits.

[0084] Browsing path: Track the sequence of products browsed by users and explore their shopping behavior patterns.

[0085] User transaction record feature extraction:

[0086] Purchase frequency: Counts the number of purchases a user makes within a specific time period.

[0087] Consumption amount: Summarize the total consumption of users and evaluate their purchasing power.

[0088] Product preferences: Based on purchase records, determine the categories of products that users frequently purchase.

[0089] Repurchase rate: Analyze users’ repeated purchase behavior for specific products or brands.

[0090] Combining features and portrait construction:

[0091] Feature fusion: Merge the extracted browsing history features with the transaction history features to form a comprehensive feature vector. For example, you can simply digitize various features and then splice them together, or use more complex feature engineering techniques such as feature crossover and dimensionality reduction to enhance the relevance and expressiveness of features.

[0092] Classification operation:

[0093] Model training: Use machine learning or deep learning algorithms (such as logistic regression, random forest, neural network, etc.) to train the above comprehensive feature vector based on existing user classification labels (such as age group, consumption level, interest group, etc.). Here, a labeled training set is required, which contains historical user data and their corresponding user portrait labels generated manually or by algorithms.

[0094] Classification operation: After training, the new user's portrait feature vector is input into the model, and the user classification label to which it belongs is predicted through classification operation. The above labels constitute the initial portrait of the user, such as "young and fashionable women", "high-end technology enthusiasts", etc., providing a basis for subsequent personalized recommendations and services.

[0095] The entire process is automatically completed through algorithms, which not only efficiently processes massive amounts of data, but also achieves a deep understanding and accurate characterization of user behavior. It is an important cornerstone of personalized services in modern e-commerce.

[0096] In one embodiment, after the step of analyzing the user behavior log based on the adjusted data analysis model to obtain the corresponding analysis result, the step further includes:

[0097] The operation data and the analysis results are associated and stored in a database.

[0098] In this embodiment, it specifically includes:

[0099] Data integration: Correlate and match the above two parts of data - operational data and user behavior analysis results. This association is based on timestamps (to ensure that operational activities and user behaviors occur in similar time periods), user IDs (to ensure that the analysis focuses on a specific user group), or other common identifiers. The purpose of association is to establish causal relationships or correlations between data in order to more fully understand how operational strategies affect user behavior and vice versa.

[0100] Storage strategy: The integrated data is stored in a database, which is a structured storage system, such as a relational database (MySQL, PostgreSQL, etc.), or a non-relational database (NoSQL, such as MongoDB, Cassandra) to adapt to complex and changing data structures. When storing, a suitable data table structure or document mode will be designed to ensure data consistency, integrity and efficient query capabilities.

[0101] Purpose and Application: The purpose of associating and storing the above data is to facilitate efficient retrieval and further analysis and utilization. Enterprises can use the above data to provide more refined operational decision support, market trend forecasting, user experience optimization, and personalized recommendation system improvement. The data in the database can be accessed by BI tools, data science teams, or automated reporting systems to quickly generate reports, dashboards, or trigger automated workflows, thereby improving the efficiency and quality of data-driven decision making.

[0102] In summary, this technical solution provides a solid foundation for enterprises' data insights and strategy formulation by integrating data from different sources but interrelated data and storing them in a structured manner.

[0103] In one embodiment, the step of differentially adjusting the model parameters of the initial data analysis model based on the model difference parameters includes:

[0104] Based on the model difference parameters, matching differentiated target model parameters from the initial data analysis model;

[0105] Based on the difference values ​​in the model difference parameters, various target model parameters are adjusted.

[0106] In this embodiment, it specifically includes:

[0107] Model difference parameters: First, we need to identify the specific parameters that cause the gap between the existing data analysis model and the desired performance or specific application scenarios. The above parameters involve model complexity, feature selection, weight distribution, learning rate, regularization strength, etc. They are directly related to the model's prediction accuracy, generalization ability, or sensitivity to specific data characteristics.

[0108] Matching target model parameters: Based on the identified model difference parameters, find out the parameters from the initial data analysis model that need special attention and adjustment. The matching process can involve algorithmic knowledge and understanding of the internal mechanism of the model to ensure that the parameters adjusted are critical and can bring about the expected changes.

[0109] Adjustment based on differential values: Once the target model parameters are determined, the above parameters are adjusted according to the specific values ​​(or guidelines) in the model differential parameters. The adjustment method is to directly set new values, increase or decrease proportionally, or apply more complex adjustment rules. For example, if the model is found to be overfitting on the training set, the model complexity may be reduced or the regularization strength may be increased; if the model generalization ability is insufficient, it is necessary to adjust the learning rate to promote better convergence, or introduce more features to capture patterns in the data.

[0110] Through such differentiated adjustments, the performance of the model on specific tasks or data sets can be effectively optimized, prediction accuracy can be improved, errors can be reduced, and generalization capabilities can be enhanced. Model parameters can be flexibly adjusted for different application scenarios and needs to make the model more suitable for actual problems and meet the needs of specific business or research. This process is often not completed in one go, but requires multiple iterations, evaluations, and adjustments to gradually approach the ideal model performance state. After each adjustment, the model is usually retrained and its performance is evaluated to ensure that the adjustment direction is correct and effective.

[0111] In short, this technical solution aims to overcome the limitations of the initial model by identifying and specifically adjusting the key parameters of the model, driving the model performance in a better direction, and better serving specific analysis goals or business scenarios.

[0112] In one embodiment, the step of obtaining the initial data analysis model includes:

[0113] Obtain the platform name information of the e-commerce platform and the latest update time of the platform;

[0114] Generate a request code based on the platform name information and the latest update time;

[0115] The request code is sent to the management end, and an initial data analysis model matching the request code is matched in the management end; wherein the management end pre-stores a mapping relationship between a plurality of different request codes and data analysis models.

[0116] In this embodiment, it specifically includes:

[0117] Platform name information: First, you need to know which e-commerce platform you are analyzing. Each platform has different requirements for data analysis models due to differences in its architecture, data structure, and business logic.

[0118] Latest update time: This information is used to identify the latest status of the platform data or structure. Because e-commerce platforms are updated regularly or irregularly, the above updates will affect the effectiveness and accuracy of data analysis, so obtaining the latest update time is crucial for selecting the most appropriate model.

[0119] Request code generation: Based on the obtained platform name information and the latest update time, a unique request code is generated. This request code is highly customized and encodes key information about the platform identity and data status to ensure the uniqueness and pertinence of the request.

[0120] Request code transmission: The generated request code is sent to a centrally managed server or cloud platform (management end). This management end maintains a database of all available data analysis models and their applicable conditions.

[0121] Mapping relationship matching: After receiving the request code, the management end will search for the matching initial data analysis model in the pre-stored mapping table. This mapping table associates different request codes with corresponding data analysis models, and each model is pre-optimized or customized for a specific platform state.

[0122] Model selection: Once a match is found, the management end will select the corresponding initial data analysis model to be deployed or directly provided to the requester. If there is no direct match, a closest model is selected according to a certain strategy (such as the most recently updated model, the default model, etc.).

[0123] Beneficial effects:

[0124] Flexibility and adaptability: The solution allows for quick adaptation to changes in the e-commerce platform, ensuring that data analysis is always based on the most appropriate model, improving the accuracy and efficiency of analysis.

[0125] Resource optimization: Centralized management of model mapping relationships reduces the work of repeated development and maintenance of similar models, while also simplifying the model deployment process and facilitating large-scale management and updates.

[0126] Standardization and automation: Through standardized request code generation and matching processes, model acquisition is automated, reducing the need for manual intervention and improving the system's response speed and processing capabilities.

[0127] In summary, this technical solution achieves efficient and accurate allocation of data analysis models for different e-commerce platforms and in different update states through a sophisticated request code mechanism and centralized management, and is an effective means to improve the quality of data analysis services.

[0128] In one embodiment, the step of generating a request code based on the platform name information and the latest update time includes:

[0129] Mapping each Chinese character in the platform name information into a corresponding coded character; wherein the platform name information is a Chinese name;

[0130] Create a blank data table with multiple rows and columns;

[0131] Add the coded characters mapped to each Chinese character to the blank data table one by one in order to obtain a character data table; wherein only one character is added to a table;

[0132] Based on the year in the latest update time, the character data table is replaced to obtain a transformation data table; based on the month and date in the latest update time, a simulation curve is generated;

[0133] The simulation curve is superimposed on the transformation data table according to a preset rule, all characters in the transformation data table located below the simulation curve are obtained, and they are combined in sequence, and the obtained character combination is used as the request code.

[0134] In this embodiment, it specifically includes:

[0135] Platform name encoding: First, map each Chinese character in the Chinese platform name to a coded character. This step can be based on Unicode or other encoding standards to ensure that each Chinese character can be uniquely identified and converted into a computer-processable form.

[0136] Create a data table: Next, build a blank data table with multiple rows and columns. This data table can be understood as a matrix, ready to store information for subsequent processing.

[0137] Fill in coded characters: fill the above coded Chinese characters into the data table one by one, and store only one character in each cell. This is actually a spatial expression of the platform name information, which is convenient for subsequent time information to affect the operation.

[0138] Year-driven data table changes: According to the year of the latest update, the character data table is "replaced" in some form. The "replacement" here can be rearrangement, rotation, replacement, etc. The specific transformation rules are determined according to the year. The purpose is to introduce changes in the time dimension so that the request code evolves over time.

[0139] Simulation curve based on month and day: Generate a simulation curve using the month and day of the latest update time. The shape, direction and other characteristics of this curve are related to the specific values ​​of the month and day, further reflecting the dynamic influence of time.

[0140] Overlay curve and filter: Overlay the simulation curve onto the transformed data table according to preset rules. Here, "overlay" can be used to determine which characters in the data table are selected based on the position of the curve. Specifically, the characters below the simulation curve are selected, and the above characters represent the final request code.

[0141] Combined characters: Finally, all the characters below the simulation curve are combined in sequence to form a string, which is the final request code. This request code not only contains direct information about the platform, but also embeds dynamic elements that reflect the platform update time, ensuring that each generated request code is unique and closely related to the current platform status.

[0142] In summary, this technical solution creatively integrates the platform name and time information through a series of carefully designed steps to generate a request code with high specificity and time-varying properties, thereby enhancing the security and effectiveness of data requests.

[0143] In one embodiment, the step of replacing the character data table based on the year in the latest update time to obtain a transformation data table; and generating a simulation curve based on the month and date in the latest update time comprises:

[0144] Obtain the year in the latest update time, and add the characters in the year in sequence to a data table with two rows and two columns to obtain a year data table; wherein the year is a four-digit number;

[0145] Superimposing the center of the year data table and the character data table, and superimposing the characters in the year data table to the overlapping positions in the character data table to obtain a transformation data table;

[0146] Obtain the month and date in the latest update time, create a coordinate system, use the number corresponding to the month as the horizontal coordinate of the feature point, and use the number corresponding to the date as the vertical coordinate of the feature point;

[0147] Based on the feature points and the origin of the coordinate system, a straight line is constructed as the simulation curve.

[0148] In this embodiment, it specifically includes:

[0149] Year characterization and layout: First, extract the four-digit year from the latest update time (for example, 2023), and then put these four digits into a two-row, two-column data table. This layout means that each digit of the year occupies a cell in the data table, forming a compact structure.

[0150] Overlay the year data table with the character data table: Next, align the year data table and overlay it onto the center of the character data table previously composed of platform name encoding characters. If there are overlapping cells between the two, the numbers in the year data table will be overlaid and combined with the original characters in the character data table. For example, in the overlapping cells, if the character in the original character data table is a and the character in the year data table is 2, then they can be combined into a2. This process ensures that the year information can significantly affect the final request code structure through the "center overlay" strategy, while also introducing the dynamic characteristics of time.

[0151] Coordinate system and feature point definition: Subsequently, the month and date are extracted from the same time information and used as the basis for constructing the simulation curve. In the two-dimensional coordinate system, the month is specified as the X-axis (horizontal axis) and the date is specified as the Y-axis (vertical axis). In this way, a specific month and date is converted into a feature point in the coordinate system.

[0152] Straight line simulation curve: Based on this characteristic point (month, date) and the origin of the coordinate system (0,0), a straight line is drawn as a simulation curve. This straight line simply and intuitively expresses a specific moment in time. Although it is called a "simulation curve", it is actually a straight line in this context because it is based only on two discrete time variables (month and date).

[0153] Through the above steps, the technical solution cleverly combines the platform name and time information (especially the year, month, and date) to generate a transformation data table and simulation curve that not only contains the platform identity but also reflects the dynamic nature of time. The transformation data table ensures the macroscopic impact of time by directly superimposing the years, while the simulation curve (although in the form of a straight line) captures the instantaneous state of time in a more abstract way. This combination not only enhances the uniqueness of the request code, but also provides an additional time dimension for security verification. It is an innovative time-sensitive security measure.

[0154] In one embodiment, the step of mapping each Chinese character in the platform name information to a corresponding coded character includes:

[0155] Obtain a standard coding table;

[0156] Perform keyword recognition on the platform name information to obtain a key name;

[0157] Rearranging the standard coding table based on the key name to obtain a rearranged coding table;

[0158] Based on the rearranged encoding table, each Chinese character in the platform name information is mapped to a corresponding encoding character.

[0159] In this embodiment, it specifically includes:

[0160] Get a standard encoding table: First, a pre-set standard encoding table is obtained. This encoding table usually contains all possible Chinese characters and their corresponding encoding characters. The encoding characters can be numbers, letters, symbols, etc., which correspond to the original Chinese characters one by one according to certain rules. For example, common encoding methods include Unicode, GBK or other custom mapping rules.

[0161] Keyword identification: Next, analyze the platform name information to identify key names. Key names can be the words in the platform name that best represent its characteristics, or the parts that require higher security. The purpose of identifying key names is to give the above characters a higher weight or special treatment, so that their role is more prominent in subsequent steps.

[0162] Rearrange the code table: Based on the identified key names, the standard code table is rearranged. The purpose of the rearrangement can be to increase the difficulty of decoding, or to optimize the encoding rules according to the characteristics of the key names. The logic of the rearrangement can be varied, such as moving the characters related to the key name to the front of the code table, or using certain attributes of the key name (such as the first letter, the number of strokes) to guide the selection order of the encoding characters, or even completely constructing the code table according to a new rule customized based on the key name.

[0163] Mapping coded characters: Finally, using the rearranged coding table, each Chinese character in the platform name information is mapped to a new coded character. This process ensures that every part of the original platform name, especially the key name, is converted into a form that is not easily directly recognizable, which improves the security level of the information. Since the coding table has been personalized according to the key name, even the same Chinese character may be mapped to different coded characters in different platform names or at different times, which increases complexity and security.

[0164] In summary, this technical solution realizes a flexible and targeted information encryption method through keyword recognition and dynamic adjustment of the coding table. It not only protects the platform name information, but also optimizes the encoding process according to specific needs, improving security while also taking into account the information processing efficiency and practicality.

[0165] Reference Figure 2 In one embodiment of the present invention, there is also provided an analysis device for online e-commerce operation data, comprising:

[0166] An acquisition unit, used to acquire operation data of the e-commerce platform; wherein the operation data includes commodity browsing records, user transaction records, and user behavior logs;

[0167] A first analysis unit, configured to analyze the commodity browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait;

[0168] A matching unit, configured to match corresponding model difference parameters in a database based on the initial user portrait; wherein a mapping relationship between the user portrait and the model difference parameters is pre-stored in the database; and the model difference parameters include differentiated model parameters and their corresponding difference values;

[0169] An adjustment unit, configured to obtain an initial data analysis model, and perform differential adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model;

[0170] The second analysis unit is used to analyze the user behavior log based on the adjusted data analysis model to obtain corresponding analysis results.

[0171] In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the above method embodiment, which will not be repeated here.

[0172] Reference Figure 3 In an embodiment of the present invention, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a display screen, an input device, a network interface and a database connected through a system bus. Among them, the processor designed by the computer 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the above method is implemented.

[0173] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0174] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0175] In summary, the analysis method, device and computer equipment for online operation data of e-commerce provided in the embodiment of the present invention include: obtaining the operation data of the e-commerce platform; wherein the operation data includes commodity browsing records, user transaction records, and user behavior logs; analyzing the commodity browsing records and user transaction records based on a preset user portrait analysis model to obtain the user's initial portrait; matching the corresponding model difference parameters in the database based on the user's initial portrait; wherein the database pre-stores a mapping relationship between the user portrait and the model difference parameters; the model difference parameters include differentiated model parameters and their corresponding difference values; obtaining the initial data analysis model, and differentially adjusting the model parameters of the initial data analysis model based on the model difference parameters to obtain the adjusted data analysis model; analyzing the user behavior log based on the adjusted data analysis model to obtain the corresponding analysis results. In the present invention, by integrating multi-source data such as user browsing records, transaction records and behavior logs, a personalized analysis model is constructed using advanced user portrait technology. The method uses a preset user portrait analysis model to deeply mine the original data and form an initial user portrait. On this basis, the concept of model difference parameters was further introduced. According to the slight differences in user characteristics, the most matching differentiated adjustment parameters were found in the database, and the basic analysis model was customized and optimized to ensure that the analysis model could better fit the characteristics of different user groups. Finally, the adjusted model was used to conduct in-depth analysis of user behavior logs to obtain more accurate and detailed analysis results, thereby helping e-commerce platforms to achieve more efficient operational decisions and personalized user experience, and improving the efficiency and accuracy of data analysis.

[0176] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and 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 provided by the present invention and used in the embodiments 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. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), 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.

[0177] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

[0178] The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for analyzing online operation data of e-commerce, characterized in that: The following steps are involved: Obtaining the operation data of the e-commerce platform; wherein the operation data includes product browsing records, user transaction records, and user behavior logs; Analyze the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait; Based on the initial user portrait, the corresponding model difference parameters are matched in the database; wherein the database pre-stores a mapping relationship between the user portrait and the model difference parameters; the model difference parameters include differentiated model parameters and their corresponding difference values; Acquire an initial data analysis model, and perform differential adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model; The user behavior log is analyzed based on the adjusted data analysis model to obtain corresponding analysis results.

2. The method for analyzing online e-commerce operation data according to claim 1, characterized in that: The step of analyzing the product browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait includes: Based on the preset user portrait analysis model, feature extraction is performed on the product browsing record and the user transaction record to obtain browsing record features and transaction record features; Combining the browsing record features with the transaction record features to obtain a portrait feature; A classification operation is performed on the portrait features to obtain a corresponding classification label as the initial portrait of the user.

3. The method for analyzing online e-commerce operation data according to claim 1, characterized in that: After the step of analyzing the user behavior log based on the adjusted data analysis model to obtain the corresponding analysis result, the method further includes: The operation data and the analysis results are associated and stored in a database.

4. The method for analyzing online e-commerce operation data according to claim 1, characterized in that: The step of differentially adjusting the model parameters of the initial data analysis model based on the model difference parameters comprises: Based on the model difference parameters, matching target model parameters with differentiation from the initial data analysis model; Based on the difference values ​​in the model difference parameters, various target model parameters are adjusted.

5. The method for analyzing online e-commerce operation data according to claim 1, characterized in that: The step of obtaining the initial data analysis model includes: Obtain the platform name information of the e-commerce platform and the latest update time of the platform; Generate a request code based on the platform name information and the latest update time; The request code is sent to the management end, and an initial data analysis model matching the request code is matched in the management end; wherein the management end pre-stores a mapping relationship between a plurality of different request codes and data analysis models.

6. The method for analyzing online e-commerce operation data according to claim 5, characterized in that: The step of generating a request code based on the platform name information and the latest update time includes: Mapping each Chinese character in the platform name information into a corresponding coded character; wherein the platform name information is a Chinese name; Create a blank data table with multiple rows and columns; Add the coded characters mapped to each Chinese character to the blank data table one by one in order to obtain a character data table; wherein only one character is added to a table; Based on the year in the latest update time, the character data table is replaced to obtain a transformation data table; based on the month and date in the latest update time, a simulation curve is generated; The simulation curve is superimposed on the transformation data table according to a preset rule, all characters in the transformation data table located below the simulation curve are obtained, and they are combined in sequence, and the obtained character combination is used as the request code.

7. The method for analyzing online e-commerce operation data according to claim 6, characterized in that: The character data table is replaced based on the year in the latest update time to obtain a transformation data table; The step of generating a simulation curve based on the month and date in the latest update time comprises: Obtain the year in the latest update time, and add the characters in the year in sequence to a data table with two rows and two columns to obtain a year data table; wherein the year is a four-digit number; Superimposing the center of the year data table and the character data table, and superimposing the characters in the year data table to the overlapping positions in the character data table to obtain a transformation data table; Obtain the month and date in the latest update time, create a coordinate system, use the number corresponding to the month as the horizontal coordinate of the feature point, and use the number corresponding to the date as the vertical coordinate of the feature point; Based on the feature points and the origin of the coordinate system, a straight line is constructed as the simulation curve.

8. The method for analyzing online e-commerce operation data according to claim 6, characterized in that: The step of mapping each Chinese character in the platform name information into a corresponding coded character comprises: Obtain a standard coding table; Perform keyword recognition on the platform name information to obtain a key name; Rearranging the standard coding table based on the key name to obtain a rearranged coding table; Based on the rearranged encoding table, each Chinese character in the platform name information is mapped to a corresponding encoding character.

9. An analysis device for online e-commerce operation data, characterized in that: include: An acquisition unit, used to acquire operation data of the e-commerce platform; wherein the operation data includes commodity browsing records, user transaction records, and user behavior logs; A first analysis unit, configured to analyze the commodity browsing records and user transaction records based on a preset user portrait analysis model to obtain an initial user portrait; A matching unit, configured to match corresponding model difference parameters in a database based on the initial user portrait; wherein a mapping relationship between the user portrait and the model difference parameters is pre-stored in the database; and the model difference parameters include differentiated model parameters and their corresponding difference values; An adjustment unit, configured to obtain an initial data analysis model, and perform differential adjustment on model parameters of the initial data analysis model based on the model difference parameters to obtain an adjusted data analysis model; The second analysis unit is used to analyze the user behavior log based on the adjusted data analysis model to obtain corresponding analysis results.

10. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images