Service data management method and system for cross-border e-commerce

Abstract

The invention relates to the technical field of data management, and discloses a cross-border e-commerce service data management method, which comprises the following steps: obtaining service data about cross-border e-commerce, and carrying out data fusion on the service data to obtain fused data; performing data expansion on the fused data to obtain expanded data, taking the expanded data as target data, and analyzing a commodity demand trend of cross-border e-commerce by using the target data; constructing a commodity supply chain network of the cross-border e-commerce, performing data supplement on the commodity supply chain network to obtain a target network, and constructing a logistics supply strategy of the cross-border e-commerce by using the target network; and querying a supply user of the cross-border e-commerce, formulating a personalized marketing strategy about the supply user, and constructing a management report of the service data based on the logistics supply strategy and the personalized marketing strategy. According to the invention, the marketing management level of the service data of cross-border e-commerce can be improved.

Description

Technical Field

[0001] The present invention relates to a service data management method and system for cross-border e-commerce, belonging to the technical field of data management. Background Art

[0002] Cross-border e-commerce services are an emerging and important form of international trade, and data management plays an important role in cross-border e-commerce services. Building a service data management method suitable for cross-border e-commerce can help e-commerce companies to easily record, track and control various service-related data, thereby improving the quality and efficiency of cross-border e-commerce services. At the same time, enterprise users can monitor the dynamic changes of business data in real time and conduct data analysis, quickly adjust service strategies, and provide more targeted services. It is also conducive to information exchange and sharing between different departments, between enterprises and partners, and between customers and enterprises, thereby reducing business risks caused by poor information flow and enhancing the efficiency of collaborative operations of enterprises.

[0003] At present, the data management method for cross-border e-commerce services usually relies on traditional data management systems to create a system with basic data entry and query functions to achieve the data management purpose of cross-border e-commerce services. However, through this traditional data management method, the important process of data analysis in data management is often ignored, making it difficult to discover the deep potential relationships in the data, and thus unable to explore the deep value of the data, resulting in a low level of marketing management of cross-border e-commerce service data. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a service data management method for cross-border e-commerce, which can improve the service data marketing management level of cross-border e-commerce.

[0005] To achieve the above-mentioned purpose, the present invention provides a cross-border e-commerce service data management method, comprising: Acquire service data about cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data; Performing data expansion on the fused data to obtain expanded data, constructing a discriminant network for the expanded data, using the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, using the expanded data as target data, and using the target data to analyze the commodity demand trend of the cross-border e-commerce; Construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain a coding vector, use the data vector and the coding vector to supplement the data of the commodity supply chain network to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce; Query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users according to the analysis results, and construct a management report on the service data based on the logistics supply strategy and the personalized marketing strategy.

[0006] Optionally, the extracting the multimodal features of the service data includes: Performing data optimization on the service data to obtain optimized data; Performing triple modal division on the optimized data to obtain triple modal data; Performing visual feature processing on the first modality data in the triple modality data to obtain a first processing feature; Performing natural language processing on the second modality data in the triple modality data to obtain a second processing feature; Performing statistical processing on the third modality data in the triple modality data to obtain a third processing feature; Based on the first processing characteristic, the second processing characteristic, and the third processing characteristic, a multimodal characteristic of the service data is determined.

[0007] Optionally, the performing data fusion on the service data based on the multimodal feature to obtain fused data includes: Performing vector conversion on the multimodal features to obtain multimodal vector features; Calculating a weighted value of the multimodal vector feature to obtain a multimodal feature weight; Based on the multimodal feature weights, the multimodal vector features are fused using the following formula to obtain fused features: in, represents the fusion feature, represents the first modality weight in the multimodal feature weight, Represents the first vector feature in the multimodal vector feature, represents the second modality weight in the multimodal feature weight, Represents the second vector feature in the multimodal vector feature, represents the third modality weight in the multimodal feature weight, Represents the third vector feature in the multimodal vector feature, represents the feature set in the first vector feature, represents the feature set in the second vector feature, represents a feature set in the third vector feature; Data reconstruction is performed on the fusion features to obtain fusion data.

[0008] Optionally, performing data expansion on the fused data to obtain expanded data includes: identifying a data expansion requirement for the fused data; Based on the data expansion requirement, construct a data expansion table of the fused data; Constructing a data generator for the fused data according to the data expansion table; The data generator is used to perform data expansion on the fused data to obtain expanded data.

[0009] Optionally, constructing a discriminant network for the expanded data includes: Constructing an initial discriminant network for the expanded data; Performing network training on the initial discriminant network to obtain a training network; The loss value of the training network is calculated using the following formula: in, represents the loss value, Indicates the number of training samples corresponding to the training network, represents the true label of the training sample corresponding to the training network, Represents the predicted label of the training sample corresponding to the training network; When the loss value meets the preset loss value, the training network is used as a discriminant network.

[0010] Optionally, the using the discriminant network to perform data authenticity assessment on the expanded data includes: Performing authenticity division on the expanded data to obtain real data and generated data; Using the discriminant network, calculating the data distribution of the real data and the generated data to obtain the real data distribution and the generated data distribution; Based on the real data distribution and the generated data distribution, the data authenticity of the expanded data is calculated using the following formula: in, Indicates the authenticity of the data, A indicates a specified data in the expanded data, Indicates that A is the real data distribution, A represents the generated data distribution, F represents the loss function, represents the differential mapping of A, Indicates the posterior probability that the specified data A is the true data, Indicates the network loss when the specified data A is the real data. Indicates the network loss when data A is specified as generated data; Based on the data authenticity, a data authenticity assessment is performed on the expanded data.

[0011] Optionally, the using the target data to analyze the commodity demand trend of the cross-border e-commerce includes: Constructing a visualization chart of the target data to visualize the target data and obtain visualization data; Performing trend analysis on the visualization data to obtain trend analysis results; Performing seasonal analysis on the visualization data to obtain seasonal analysis results; Performing periodic analysis on the visualization data to obtain periodic analysis results; Based on the trend analysis results, the seasonal analysis results and the periodic analysis results, the commodity demand trend of the cross-border e-commerce is determined.

[0012] Optionally, the utilizing the target network to construct the logistics supply strategy of the cross-border e-commerce includes: Identify supply chain decision parameters of the cross-border e-commerce company according to the target network; Performing gene sequence encoding on the supply chain decision parameters to obtain gene sequence parameters; Randomly configuring the gene sequence parameters to obtain an initial parameter sequence; evaluating the sequence fitness of the initial parameter sequence; Extracting an initial parameter sequence corresponding to a preferred fitness in the sequence fitness to obtain a target parameter sequence; Performing crossover mutation on the target parameter sequence to obtain an updated parameter sequence; Extracting a parameter sequence with the best fitness from the update parameter sequence to obtain a final parameter sequence; Based on the decision parameters corresponding to the final parameter sequence, a logistics supply strategy for the cross-border e-commerce is constructed.

[0013] Optionally, formulating a personalized marketing strategy for the supply user according to the analysis result includes: Building a product recommendation strategy based on the user preference analysis results, purchase history analysis results, and hot trend analysis results in the analysis results; Building a product pricing strategy based on the differentiated pricing analysis results and the dynamic pricing analysis results in the analysis results; Building a product promotion strategy based on the promotion activity analysis results and the holiday promotion analysis results in the analysis results; Based on the product recommendation strategy, the product pricing strategy and the product price strategy, a personalized marketing strategy for the supply users is formulated.

[0014] In a second aspect, the present invention provides a cross-border e-commerce service data management system, the system comprising: A data fusion module is used to obtain service data about cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data; A commodity demand trend analysis module is used to perform data expansion on the fused data to obtain expanded data, construct a discriminant network for the expanded data, use the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, use the expanded data as target data, and use the target data to analyze the commodity demand trend of the cross-border e-commerce; A logistics supply strategy construction module is used to construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain an encoding vector, use the data vector and the encoding vector to supplement the commodity supply chain network with data to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce; A management report generation module is used to query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users based on the analysis results, and construct a management report on the service data based on the logistics supply strategy and the personalized marketing strategy.

[0015] Compared with the problems described in the background technology, the present invention first obtains service data about cross-border e-commerce and extracts multimodal features of the service data to comprehensively interpret the data from different angles, thereby obtaining richer data information. Among them, the multimodal features refer to a set of features extracted from multiple different modes, such as image, numerical value, and text data modes, and then the service data is subjected to data fusion to obtain fused data that can integrate the features of different modes, give full play to the advantages of each modality data, and improve the accuracy and comprehensiveness of the analysis; further, the embodiment of the present invention expands the fused data to obtain expanded data to increase the data volume of service data in cross-border e-commerce, avoid overfitting of data analysis caused by limited data in cross-border e-commerce (in cross-border e-commerce, the amount of data is often limited, especially for some emerging commodities or markets), and after analyzing the authenticity of the expanded data, the expanded data is used to analyze the commodity demand trend of the cross-border e-commerce to help cross-border e-commerce enterprises to carry out inventory management, procurement decisions, market promotion, etc., to improve the operational efficiency and competitiveness of enterprises; further, the present invention can better manage by constructing a commodity supply chain network of cross-border e-commerce. Logistics and inventory, improve the efficiency and reliability of the supply chain, and by utilizing the target network, construct the logistics supply strategy of the cross-border e-commerce, it is possible to design a logistics supply strategy suitable for cross-border e-commerce, thereby optimizing the logistics process, improving logistics efficiency, reducing logistics costs, and improving the overall operation level and customer satisfaction of cross-border e-commerce; the present invention analyzes the consumption behavior of supply users, and the analysis results can deeply understand the user's demand preferences, purchasing habits, consumer psychology, etc., thereby providing a basis for formulating personalized marketing strategies, and according to the analysis results, formulate personalized marketing strategies for supply users to provide products and services that meet user expectations, improve user satisfaction, and based on the logistics supply strategy and the personalized marketing strategy, construct a management report on the service data to monitor the execution of logistics supply and personalized marketing in real time, evaluate the effectiveness and implementation effect of the strategy, promptly discover problems and make adjustments, thereby improving the marketing management level of service data in cross-border e-commerce. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a process flow of a cross-border e-commerce service data management method provided by an embodiment of the present invention; Figure 2 A module diagram of a cross-border e-commerce service data management system provided by one embodiment of the present invention.

[0017] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0019] The embodiment of the present application provides a service data management method for cross-border e-commerce. The execution subject of the service data management method for cross-border e-commerce includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the method provided by the embodiment of the present application. In other words, the service data management method for cross-border e-commerce can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0020] Embodiment 1: Reference Figure 1 FIG. 1 is a flow chart of a cross-border e-commerce service data management method provided by an embodiment of the present invention. In this embodiment, the cross-border e-commerce service data management method includes: S1. Obtain service data on cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data.

[0021] The embodiment of the present invention can obtain basic data for in-depth analysis and effective management by acquiring service data about cross-border e-commerce.

[0022] It should be noted that in the field of cross-border e-commerce, service data covers multiple aspects, including user transaction records, browsing behavior, evaluation feedback, logistics information and other data.

[0023] Optionally, the service data can be obtained through channels such as the transaction system and user logs of the cross-border e-commerce platform.

[0024] Furthermore, the embodiments of the present invention can comprehensively interpret the data from different angles by extracting the multimodal features of the service data, thereby obtaining richer data information. The multimodal features refer to a set of features extracted from multiple different modalities, such as image, numerical value, and text data modalities.

[0025] As an embodiment of the present invention, the extracting of multimodal features of the service data includes: performing data optimization on the service data to obtain optimized data, performing triply modal division on the optimized data to obtain triply modal data, performing visual feature processing on a first modal data in the triply modal data to obtain a first processing feature, performing natural language processing on a second modal data in the triply modal data to obtain a second processing feature, performing statistical processing on a third modal data in the triply modal data to obtain a third processing feature, and determining the multimodal features of the service data based on the first processing feature, the second processing feature, and the third processing feature.

[0026] Among them, the optimized data refers to data obtained by data preprocessing before analyzing the data, the first modality data refers to image data, the second modality data refers to text data, and the third modality data refers to numerical data.

[0027] Optionally, the optimized data can be obtained by performing deduplication on the service data using a hash algorithm and removing abnormal data using an interpolation algorithm, the triple modal data can be obtained by performing data classification on the optimized data using a decision tree algorithm, the first processing feature can be obtained by performing image feature processing on the first modality data using an image processing algorithm of a neural network, the second processing feature can be obtained by performing text processing on the second modality data using a word vector model, and the third processing feature can be directly obtained from a database.

[0028] Furthermore, the embodiment of the present invention performs data fusion on the service data based on the multimodal features to obtain fused data, which can integrate the features of different modalities, give full play to the advantages of each modal data, and improve the accuracy and comprehensiveness of the analysis.

[0029] As an embodiment of the present invention, the data fusion of the service data based on the multimodal features to obtain fused data includes: performing vector conversion on the multimodal features to obtain multimodal vector features, calculating weighted values ​​of the multimodal vector features to obtain multimodal feature weights, and based on the multimodal feature weights, performing feature fusion on the multimodal vector features using the following formula to obtain fused features: in, represents the fusion feature, represents the first modality weight in the multimodal feature weight, Represents the first vector feature in the multimodal vector feature, represents the second modality weight in the multimodal feature weight, Represents the second vector feature in the multimodal vector feature, represents the third modality weight in the multimodal feature weight, Represents the third vector feature in the multimodal vector feature, represents the feature set in the first vector feature, represents the feature set in the second vector feature, represents a feature set in the third vector feature; Data reconstruction is performed on the fusion features to obtain fusion data.

[0030] Optionally, the multimodal vector features can be obtained by transforming multimodal features through a digital matrix, the multimodal feature weights can be calculated through an automatic weight learning algorithm in machine learning, and the fused data can be obtained by using a machine learning algorithm, such as a support vector machine, to annotate the fused features and then perform data classification.

[0031] S2. Perform data expansion on the fused data to obtain expanded data, construct a discriminant network for the expanded data, use the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, use the expanded data as target data, and use the target data to analyze the commodity demand trend of the cross-border e-commerce.

[0032] The embodiment of the present invention performs data expansion on the fused data to obtain expanded data, which can increase the data volume of service data in cross-border e-commerce and avoid overfitting in data analysis due to limited data in cross-border e-commerce (in cross-border e-commerce, the data volume is often limited, especially for some emerging commodities or markets).

[0033] As an embodiment of the present invention, the data expansion of the fused data to obtain the expanded data includes: identifying the data expansion requirements of the fused data, constructing a data expansion table of the fused data based on the data expansion requirements, constructing a data generator of the fused data according to the data expansion table, and using the data generator to perform data expansion on the fused data to obtain the expanded data.

[0034] The data expansion table refers to a table tool used to clarify specific requirements for data expansion.

[0035] Optionally, the data expansion demand can be determined by analyzing the characteristics and application scenarios of cross-border e-commerce service data to determine the requirements for the type, quantity, characteristics, etc. of the data that need to be expanded. For example, in cross-border e-commerce, if it is found that there is less data for certain product categories, it may be necessary to expand the data for these categories to improve the generalization ability and accuracy of the model. The data expansion table can be constructed through an Excel spreadsheet tool according to the data expansion demand to clearly specify the type of data that needs to be expanded, such as product category, required data characteristics (such as price range, sales distribution, etc.) and expansion quantity target. The data generator can be constructed through a generative adversarial network in combination with the data expansion table.

[0036] Exemplarily, the process of using the data generator to perform data expansion on the fused data to obtain the expanded data is as follows: for the image data in the fused data, a deep convolutional generative adversarial network (DCGAN) can be used, which consists of a convolutional layer, a deconvolutional layer, a batch normalization layer, etc., to generate realistic product images; for the text data in the fused data, a generative adversarial network based on a recurrent neural network (RNN) or a long short-term memory network (LSTM) can be used to generate natural and fluent user review texts; for the sales data in the fused data, a fully connected neural network can be used to generate a numerical sequence similar to the real sales data.

[0037] The embodiment of the present invention can obtain an analysis tool that can quickly analyze the accuracy of data by constructing the discriminant network of the expanded data, thereby improving the efficiency and authenticity of data analysis. Wherein, the discriminant network refers to a machine learning model that is trained to distinguish between real data and expanded data.

[0038] As an embodiment of the present invention, the construction of the discriminant network of the expanded data includes: constructing an initial discriminant network of the expanded data, performing network training on the initial discriminant network to obtain a training network, and calculating the loss value of the training network using the following formula: in, represents the loss value, Indicates the number of training samples corresponding to the training network, represents the true label of the training sample corresponding to the training network, Represents the predicted label of the training sample corresponding to the training network; When the loss value meets the preset loss value, the training network is used as a discriminant network.

[0039] Optionally, the initial discriminant network can be constructed using the GAN architecture in deep learning, the training network can be trained by collecting historical service data of cross-border e-commerce, and the preset loss value can be set through business indicators of cross-border e-commerce. For example, business indicators such as product sales growth rate and user satisfaction improvement can be used as part of the loss function, or a specific loss threshold can be set according to business needs. When the network reaches or exceeds this threshold in terms of business indicators, it is considered that the loss value meets the requirements. For example, if the business goal is to increase the product sales growth rate by 10%, the difference between the sales growth rate predicted by the model and the actual sales growth rate can be used as the loss function, and a loss threshold can be set. When the loss value is less than this threshold, it is considered that the model meets the business needs. The specific setting needs to be made in combination with the actual application scenario.

[0040] Furthermore, the embodiment of the present invention can ensure the quality of the expanded data by using the discriminant network to perform data authenticity evaluation on the expanded data, avoid introducing false or low-quality data, and thus ensure the reliability of the analysis results.

[0041] As an embodiment of the present invention, the use of the discriminant network to perform data authenticity assessment on the expanded data includes: performing authenticity division on the expanded data to obtain real data and generated data, using the discriminant network to calculate the data distribution of the real data and the generated data to obtain real data distribution and generated data distribution, and based on the real data distribution and the generated data distribution, using the following formula to calculate the data authenticity of the expanded data: in, Indicates the authenticity of the data, A indicates a specified data in the expanded data, Indicates that A is the real data distribution, A represents the generated data distribution, F represents the loss function, represents the differential mapping of A, Indicates the posterior probability that the specified data A is the true data, Indicates the network loss when the specified data A is the real data. Indicates the network loss when data A is specified as generated data; Based on the data authenticity, a data authenticity assessment is performed on the expanded data.

[0042] Optionally, the expanded data is divided into realism, and the real data and generated data can be preliminarily divided according to the source of the data, the known original real data is marked as real data, and the data obtained by methods such as data expansion is marked as generated data. The discriminant network is used to calculate the data distribution of the real data and the generated data, and the real data distribution and the generated data distribution can be obtained by inputting the real data into the discriminant network and statistically analyzing the distribution of the real data in different feature dimensions through the output of the network. For example, if the discriminant network is a deep neural network that processes product image data, then the distribution of the real product image data can be estimated by analyzing the feature representation of the middle layer of the network or the probability distribution of the output layer. Similarly, the generated data is input into the discriminant network to obtain the distribution of the generated data. In cross-border e-commerce, for example, for the generated data of commodity sales data, the generated data distribution can be determined by analyzing the distribution of generated sales quantity, sales time and other characteristics. The data authenticity assessment of the expanded data based on the data authenticity can be performed by comparing the calculated data authenticity value with a preset authenticity threshold. If the data authenticity value is not less than the preset authenticity threshold, it indicates that the corresponding data is credible data. If the data authenticity value is less than the preset authenticity threshold, it indicates that the corresponding data is untrustworthy data.

[0043] It should be explained that when the evaluation result of the data authenticity evaluation is excellent, taking the expanded data as the target data means that when the calculation result of the data authenticity is not less than 0.8 (that is, the preset authenticity threshold), reliable analysis data can be obtained, thereby ensuring the accuracy of subsequent analysis results. If the evaluation result of the data authenticity evaluation is not excellent, it is necessary to return to the step of expanding the fused data to obtain the expanded data until the evaluation result of the data authenticity evaluation is excellent. It should be further explained that the preset authenticity threshold needs to be adaptively set in combination with the actual application scenario.

[0044] Furthermore, the embodiments of the present invention can help cross-border e-commerce companies to perform inventory management, procurement decisions, market promotion, etc. by utilizing the target data to analyze the commodity demand trend of the cross-border e-commerce, thereby improving the operational efficiency and competitiveness of the companies.

[0045] As an embodiment of the present invention, the use of the target data to analyze the commodity demand trend of the cross-border e-commerce includes: constructing a visualization chart of the target data to visualize the target data to obtain visualization data, performing trend analysis on the visualization data to obtain trend analysis results, performing seasonal analysis on the visualization data to obtain seasonal analysis results, performing periodic analysis on the visualization data to obtain periodic analysis results, and determining the commodity demand trend of the cross-border e-commerce based on the trend analysis results, the seasonal analysis results and the periodic analysis results.

[0046] The visualization chart refers to a tool that displays data in a graphical manner.

[0047] Optionally, the visualization chart can be constructed using the programming language Python and a data visualization library. The trend analysis results can be obtained by observing the visualization data using a line chart, and observing whether the overall discount of the line chart is rising, falling, or remaining stable. The seasonal analysis results can be obtained by representing the visualization data using a time series line chart, and observing whether the time series line chart has a regular repeating pattern. For example, sales growth during specific holidays or promotional activities indicates that there may be seasonal fluctuations. These special time periods can be marked and the data changes in the same time periods in different years can be compared. The periodic analysis results can be obtained by observing whether the time series line chart has longer-term fluctuations. The commodity demand trend can be obtained by calculating the statistical characteristics corresponding to the trend analysis results, the seasonal analysis results, and the periodic analysis results.

[0048] S3. Construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain a coding vector, use the data vector and the coding vector to supplement the data of the commodity supply chain network to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce.

[0049] The embodiment of the present invention can better manage logistics and inventory and improve the efficiency and reliability of the supply chain by constructing the cross-border e-commerce commodity supply chain network. The commodity supply chain network refers to a complex network system composed of relevant entities and their activities involved in various links such as commodity production, circulation and sales.

[0050] Optionally, the commodity supply chain network can be constructed by imitating the structure of a biological neural network to establish a supply chain network model. The nodes in the supply chain network model can represent each link in the supply chain, such as suppliers, warehouses, transportation nodes, sales points, etc. The connection weights can represent the strength of the relationship between each link, such as the frequency of logistics transportation, the amount of cargo flow, etc.

[0051] The example of the present invention identifies the logistics data in the target data and converts the logistics data into vectors to obtain data vectors that can accurately focus on the data of the logistics link, provide special information support for subsequent logistics strategy formulation, and make the data more normalized and standardized through vector conversion, which is convenient for mathematical calculations and model analysis, and can more efficiently mine the potential rules and characteristics in the logistics data.

[0052] Optionally, the logistics data may be obtained by querying a corresponding logistics data storage node in the corresponding database, and the data vector may be obtained by converting the logistics data using a script constructed using Java.

[0053] The embodiment of the present invention can convert the abstract commodity demand trend into a specific digital coding form by quantizing and encoding the commodity demand trend to obtain a coding vector, so that the commodity demand trend becomes measurable and comparable, and the changes in market demand can be understood more intuitively.

[0054] Optionally, the encoding vector may be obtained by performing dimensionality reduction processing on multiple demand-related variables in the commodity demand trend through principal component analysis technology, extracting main demand trend components, and then quantizing and encoding these components.

[0055] Furthermore, the embodiment of the present invention utilizes the data vector and the coding vector to supplement the data of the commodity supply chain network to obtain the target network, which can make the supply chain network richer and more comprehensive. It not only includes information on traditional supply, production, sales and other links, but also adds dynamic data on logistics and demand trends, thereby being able to formulate more accurate and effective logistics supply strategies.

[0056] Furthermore, the embodiments of the present invention can design a logistics supply strategy suitable for cross-border e-commerce by utilizing the target network to construct the logistics supply strategy of the cross-border e-commerce, thereby optimizing the logistics process, improving logistics efficiency, reducing logistics costs, and improving the overall operation level and customer satisfaction of the cross-border e-commerce.

[0057] As an embodiment of the present invention, the use of the target network to construct the logistics supply strategy of the cross-border e-commerce includes: identifying the supply chain decision parameters of the cross-border e-commerce according to the target network, performing gene sequence encoding on the supply chain decision parameters to obtain gene sequence parameters, randomly configuring the gene sequence parameters to obtain an initial parameter sequence, evaluating the sequence fitness of the initial parameter sequence, extracting the initial parameter sequence corresponding to the preferred fitness in the sequence fitness to obtain a target parameter sequence, performing cross-mutation on the target parameter sequence to obtain an updated parameter sequence, extracting a parameter sequence with the best fitness in the updated parameter sequence to obtain a final parameter sequence, and constructing the logistics supply strategy of the cross-border e-commerce based on the decision parameters corresponding to the final parameter sequence.

[0058] The supply chain decision parameters refer to key variables and factors used to make decisions and optimize supply chain operations in supply chain management, such as inventory levels, logistics path selection, and supplier combinations.

[0059] Optionally, the supply chain decision parameters can be determined by extracting various data and information related to cross-border e-commerce logistics supply from the target network. For example, from the logistics data part of the target network, transportation cost, transportation time, node information of different logistics paths, etc. can be identified as decision parameters. The gene sequence parameters can be implemented by binary encoding. The initial parameter sequence can be randomly combined by using a random number generator or a specific random algorithm to generate a set of initial supply chain configurations. For example: assuming that there are three encodings for inventory levels, five different possible logistics paths, and four supplier combinations, then randomly extract combinations from these options to generate an initial parameter sequence, such as "medium inventory (01) + path D (encoded as 'D') + supplier combination 2 (encoded as '23')". The sequence fitness can be determined by determining the fitness function, taking into account multiple factors such as market trend prediction results, cost, efficiency, and service level. For example, the fitness function can be set as the ratio of market demand satisfaction rate to total cost. The market demand satisfaction rate can be evaluated based on the commodity demand trend in the target network and information such as inventory level and logistics path in the initial parameter sequence. If the inventory level can better meet the market demand and the logistics path can deliver the goods in time, then the market demand satisfaction rate is high. The total cost includes transportation cost, warehousing cost, inventory cost, etc.The corresponding cost is calculated according to the mode of transportation, inventory level, etc. in the initial parameter sequence. Finally, the fitness value of each initial parameter sequence is calculated according to the fitness function. The target parameter sequence can sort the initial parameter sequence from high to low according to the fitness value, and select a part of the sequences with higher fitness as the target parameter sequence. For example, the initial parameter sequences ranked in the top 20% of fitness are selected as the target parameter sequence. These sequences represent relatively better supply chain configurations under the current random configuration. The updated parameter sequence can be obtained by randomly selecting two target parameter sequences and exchanging genes according to certain crossover rules. For example, for logistics path selection genes and inventory level genes, crossover can be performed at the middle position of the two sequences. If one sequence is "high inventory (10) + path E (coded as 'E') + supplier combination 1 (coded as '12')", and the other sequence is "medium inventory (01) + path F (coded as 'F') + supplier combination 3 (coded as '24')", after crossover at the path gene position, a new sequence "high inventory (10) + path F (coded as 'F') + Supplier combination 1 (coded as '12')" and "medium inventory (01) + path E (coded as 'E') + supplier combination 3 (coded as '24')", the final parameter sequence can be evaluated by using the fitness function again to update the parameter sequence, and the parameter sequence with the highest fitness is selected as the final parameter sequence. The decision parameters corresponding to the final parameter sequence are used to construct the logistics supply strategy of the cross-border e-commerce. The specific logistics supply strategy can be formulated according to the actual decision parameters corresponding to the encoding in the final parameter sequence. For example, the final parameter sequence is represented as "low inventory (00) + path G (coded as 'G') + supplier combination 4 (coded as '13')", then the corresponding logistics supply strategy can be: maintain a low inventory level to reduce costs, select path G for cargo transportation, and cooperate with suppliers in supplier combination 4 to ensure a stable supply of goods. At the same time, specific inventory management plans, transportation plans and supplier cooperation strategies are formulated according to these decision parameters to achieve efficient logistics supply of cross-border e-commerce.

[0060] S4. Query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users according to the analysis results, and construct a management report of the service data based on the logistics supply strategy and the personalized marketing strategy.

[0061] The embodiment of the present invention can clarify the target group by querying the supply users of the cross-border e-commerce, and then formulate a targeted marketing strategy to improve the utilization efficiency of resources.

[0062] Optionally, the supply user can be obtained by querying the backend management system of the cross-border e-commerce platform.

[0063] The embodiment of the present invention performs consumption behavior analysis on the supply user, and the analysis result can provide an in-depth understanding of the user's demand preferences, purchasing habits, consumption psychology, etc., thereby providing a basis for formulating personalized marketing strategies.

[0064] Optionally, the consumption behavior analysis of the supply user can be obtained by analyzing the purchase behavior, product preference, functional requirements, and channel preference of the supply user.

[0065] Furthermore, the embodiment of the present invention can provide products and services that meet user expectations and improve user satisfaction by formulating a personalized marketing strategy for the supply user based on the analysis results.

[0066] As an embodiment of the present invention, the personalized marketing strategy for the supply user is formulated based on the analysis results, including: constructing a product recommendation strategy based on the user preference analysis results, purchase history analysis results and hot trend analysis results in the analysis results, constructing a product price strategy based on the differentiated pricing analysis results and dynamic pricing analysis results in the analysis results, constructing a product promotion strategy based on the promotion activity analysis results and holiday promotion analysis results in the analysis results, and formulating a personalized marketing strategy for the supply user based on the product recommendation strategy, the product price strategy and the product price strategy.

[0067] Among them, the user preference analysis result refers to the result obtained by analyzing the user's product purchase preference, the purchase history analysis result refers to the result obtained by analyzing the user's product purchase history, the hot trend analysis result refers to the result obtained by combining the market's hot trends and user's needs analysis, the differentiated pricing analysis result refers to the result obtained by analyzing the user's consumption level and purchase history, the dynamic pricing analysis result refers to the result obtained by analyzing the market's supply and demand situation and the user's purchasing behavior, the promotion activity analysis result refers to the personalized promotion activity plan formulated according to the user's interests and purchase history, and the holiday promotion analysis result refers to the promotion plan launched in combination with different holidays and anniversaries.

[0068] Furthermore, the embodiment of the present invention can monitor the execution of logistics supply and personalized marketing in real time, evaluate the effectiveness and implementation effect of the strategy, discover problems in time and make adjustments, thereby improving the marketing management level of service data in cross-border e-commerce by constructing the management report of the service data based on the logistics supply strategy and the personalized marketing strategy.

[0069] Compared with the problems described in the background technology, the present invention first obtains service data about cross-border e-commerce and extracts multimodal features of the service data to comprehensively interpret the data from different angles, thereby obtaining richer data information. Among them, the multimodal features refer to a set of features extracted from multiple different modes, such as image, numerical value, and text data modes, and then the service data is subjected to data fusion to obtain fused data that can integrate the features of different modes, give full play to the advantages of each modality data, and improve the accuracy and comprehensiveness of the analysis; further, the embodiment of the present invention expands the fused data to obtain expanded data to increase the data volume of service data in cross-border e-commerce, avoid overfitting of data analysis caused by limited data in cross-border e-commerce (in cross-border e-commerce, the amount of data is often limited, especially for some emerging commodities or markets), and after analyzing the authenticity of the expanded data, the expanded data is used to analyze the commodity demand trend of the cross-border e-commerce to help cross-border e-commerce enterprises to carry out inventory management, procurement decisions, market promotion, etc., to improve the operational efficiency and competitiveness of enterprises; further, the present invention can better manage by constructing a commodity supply chain network of cross-border e-commerce. Logistics and inventory, improve the efficiency and reliability of the supply chain, and by utilizing the target network, construct the logistics supply strategy of the cross-border e-commerce, it is possible to design a logistics supply strategy suitable for cross-border e-commerce, thereby optimizing the logistics process, improving logistics efficiency, reducing logistics costs, and improving the overall operation level and customer satisfaction of cross-border e-commerce; the present invention analyzes the consumption behavior of supply users, and the analysis results can deeply understand the user's demand preferences, purchasing habits, consumer psychology, etc., thereby providing a basis for formulating personalized marketing strategies, and according to the analysis results, formulate personalized marketing strategies for supply users to provide products and services that meet user expectations, improve user satisfaction, and based on the logistics supply strategy and the personalized marketing strategy, construct a management report on the service data to monitor the execution of logistics supply and personalized marketing in real time, evaluate the effectiveness and implementation effect of the strategy, promptly discover problems and make adjustments, thereby improving the marketing management level of service data in cross-border e-commerce.

[0070] Embodiment 2: like Figure 2 , is a system functional module diagram of a method for implementing service data management for cross-border e-commerce provided by an embodiment of the present invention.

[0071] The cross-border e-commerce service data management system 200 of the present invention can be installed in an electronic device. According to the functions implemented, the cross-border e-commerce service data management system can include a data fusion module 201, a commodity demand trend analysis module 202, a logistics supply strategy construction module 203, and a management report generation module 204. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0072] In the embodiment of the present invention, the functions of each module / unit are as follows: The data fusion module 201 is used to obtain service data about cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data; The commodity demand trend analysis module 202 is used to perform data expansion on the fused data to obtain expanded data, build a discriminant network for the expanded data, use the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, use the expanded data as target data, and use the target data to analyze the commodity demand trend of the cross-border e-commerce; The logistics supply strategy construction module 203 is used to construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain a coding vector, use the data vector and the coding vector to supplement the data of the commodity supply chain network to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce; The management report generation module 204 is used to query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users based on the analysis results, and construct a management report on the service data based on the logistics supply strategy and the personalized marketing strategy.

[0073] In detail, each module in the cross-border e-commerce service data management system 200 described in the embodiment of the present invention adopts the same method as above when used. Figure 1 The technical means are the same as the cross-border e-commerce service data management method described in, and can produce the same technical effects, so I will not go into details here.

[0074] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0075] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0077] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0078] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0079] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A cross-border e-commerce service data management method, characterized in that: The method comprises: Acquire service data about cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data; Performing data expansion on the fused data to obtain expanded data, constructing a discriminant network for the expanded data, using the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, using the expanded data as target data, and using the target data to analyze the commodity demand trend of the cross-border e-commerce; Construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain a coding vector, use the data vector and the coding vector to supplement the data of the commodity supply chain network to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce; Query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users according to the analysis results, and construct a management report on the service data based on the logistics supply strategy and the personalized marketing strategy.

2. The cross-border e-commerce service data management method according to claim 1, characterized in that: The extracting the multimodal features of the service data includes: Performing data optimization on the service data to obtain optimized data; Performing triple modal division on the optimized data to obtain triple modal data; Performing visual feature processing on the first modality data in the triple modality data to obtain a first processed feature; Performing natural language processing on the second modality data in the triple modality data to obtain a second processing feature; Performing statistical processing on the third modality data in the triple modality data to obtain a third processing feature; Based on the first processing characteristic, the second processing characteristic, and the third processing characteristic, a multimodal characteristic of the service data is determined.

3. The cross-border e-commerce service data management method according to claim 1, characterized in that: The performing data fusion on the service data based on the multimodal features to obtain fused data includes: Performing vector conversion on the multimodal features to obtain multimodal vector features; Calculating a weighted value of the multimodal vector feature to obtain a multimodal feature weight; Based on the multimodal feature weights, the multimodal vector features are fused using the following formula to obtain fused features: in, represents the fusion feature, represents the first modality weight in the multimodal feature weight, Represents the first vector feature in the multimodal vector feature, represents the second modality weight in the multimodal feature weight, Represents the second vector feature in the multimodal vector feature, represents the third modality weight in the multimodal feature weight, Represents the third vector feature in the multimodal vector feature, represents the feature set in the first vector feature, represents the feature set in the second vector feature, represents a feature set in the third vector feature; Data reconstruction is performed on the fusion features to obtain fusion data.

4. The cross-border e-commerce service data management method according to claim 1, characterized in that: The step of performing data expansion on the fused data to obtain expanded data includes: identifying a data expansion requirement for the fused data; Based on the data expansion requirement, construct a data expansion table of the fused data; Constructing a data generator for the fused data according to the data expansion table; The data generator is used to perform data expansion on the fused data to obtain expanded data.

5. The cross-border e-commerce service data management method according to claim 1, characterized in that: The step of constructing a discriminant network for the expanded data includes: Constructing an initial discriminant network for the expanded data; Performing network training on the initial discriminant network to obtain a training network; The loss value of the training network is calculated using the following formula: in, represents the loss value, Indicates the number of training samples corresponding to the training network, represents the true label of the training sample corresponding to the training network, Represents the predicted label of the training sample corresponding to the training network; When the loss value meets the preset loss value, the training network is used as a discriminant network.

6. The cross-border e-commerce service data management method according to claim 1, characterized in that: The using the discriminant network to perform data authenticity evaluation on the expanded data includes: Performing authenticity division on the expanded data to obtain real data and generated data; Using the discriminant network, calculating the data distribution of the real data and the generated data to obtain the real data distribution and the generated data distribution; Based on the real data distribution and the generated data distribution, the data authenticity of the expanded data is calculated using the following formula: in, Indicates the authenticity of the data, A indicates a specified data in the expanded data, Indicates that A is the real data distribution, A represents the generated data distribution, F represents the loss function, represents the differential mapping of A, Indicates the posterior probability that the specified data A is the true data, Indicates the network loss when the specified data A is the real data. Indicates the network loss when data A is specified as generated data; Based on the data authenticity, a data authenticity assessment is performed on the expanded data.

7. The cross-border e-commerce service data management method according to claim 1, characterized in that: The using the target data to analyze the commodity demand trend of the cross-border e-commerce includes: Constructing a visualization chart of the target data to visualize the target data and obtain visualization data; Performing trend analysis on the visualization data to obtain trend analysis results; Performing seasonal analysis on the visualization data to obtain seasonal analysis results; Performing periodic analysis on the visualization data to obtain periodic analysis results; Based on the trend analysis results, the seasonal analysis results and the periodic analysis results, the commodity demand trend of the cross-border e-commerce is determined.

8. The cross-border e-commerce service data management method according to claim 1, characterized in that: The use of the target network to construct the logistics supply strategy of the cross-border e-commerce includes: Identify supply chain decision parameters of the cross-border e-commerce company according to the target network; Performing gene sequence encoding on the supply chain decision parameters to obtain gene sequence parameters; Randomly configuring the gene sequence parameters to obtain an initial parameter sequence; evaluating the sequence fitness of the initial parameter sequence; Extracting an initial parameter sequence corresponding to a preferred fitness in the sequence fitness to obtain a target parameter sequence; Performing crossover mutation on the target parameter sequence to obtain an updated parameter sequence; Extracting a parameter sequence with the best fitness from the update parameter sequence to obtain a final parameter sequence; Based on the decision parameters corresponding to the final parameter sequence, a logistics supply strategy for the cross-border e-commerce is constructed.

9. The cross-border e-commerce service data management method according to claim 1, characterized in that: The step of formulating a personalized marketing strategy for the supply user according to the analysis result includes: Building a product recommendation strategy based on the user preference analysis results, purchase history analysis results, and hot trend analysis results in the analysis results; Building a product pricing strategy based on the differentiated pricing analysis results and the dynamic pricing analysis results in the analysis results; Building a product promotion strategy based on the promotion activity analysis results and the holiday promotion analysis results in the analysis results; Based on the product recommendation strategy, the product pricing strategy and the product price strategy, a personalized marketing strategy for the supply users is formulated.

10. A cross-border e-commerce service data management system, characterized in that: The system comprises: A data fusion module is used to obtain service data about cross-border e-commerce, extract multimodal features of the service data, and perform data fusion on the service data based on the multimodal features to obtain fused data; A commodity demand trend analysis module is used to perform data expansion on the fused data to obtain expanded data, construct a discriminant network for the expanded data, use the discriminant network to perform data authenticity evaluation on the expanded data, and when the evaluation result of the data authenticity evaluation is excellent, use the expanded data as target data, and use the target data to analyze the commodity demand trend of the cross-border e-commerce; A logistics supply strategy construction module is used to construct the commodity supply chain network of the cross-border e-commerce, identify the logistics data in the target data, and perform vector conversion on the logistics data to obtain a data vector, quantify and encode the commodity demand trend to obtain an encoding vector, use the data vector and the encoding vector to supplement the commodity supply chain network with data to obtain a target network, and use the target network to construct the logistics supply strategy of the cross-border e-commerce; A management report generation module is used to query the supply users of the cross-border e-commerce, analyze the consumption behavior of the supply users, obtain analysis results, formulate personalized marketing strategies for the supply users based on the analysis results, and construct a management report on the service data based on the logistics supply strategy and the personalized marketing strategy.

Images