An internet-of-things coding-based cross-border e-commerce user portrait data processing method

Abstract

This invention discloses a method for processing cross-border e-commerce user profile data based on IoT coding, comprising the following steps: collecting data from a cross-border e-commerce platform to generate a multi-source behavior dataset; generating a unique IoT coding identifier and establishing a set of coded objects; constructing a user behavior relationship set based on the interaction relationships between the coded objects; constructing a coding behavior relationship matrix to generate an initial behavior trajectory space; performing a first-layer spatial folding process to generate a unified behavior trajectory space; generating behavior structure parameters and performing a second-layer spatial folding process to generate a double-layer folded behavior trajectory space structure; extracting user behavior features to generate a user behavior feature set; and constructing a user profile data structure to generate cross-border e-commerce user profile data. This invention employs an IoT coding system and a double-layer spatial folding process to accurately construct cross-border e-commerce user profiles, possessing advantages such as strong data integration, high behavioral correlation, and high profile accuracy.

Description

Technical Field

[0001] This invention relates to the field of cross-border e-commerce user profile data processing, and in particular to a method for processing cross-border e-commerce user profile data based on Internet of Things (IoT) coding. Background Technology

[0002] As cross-border e-commerce platforms continue to expand, user behavior data on these platforms is becoming increasingly multi-sourced and dispersed. This data is typically stored across multiple business systems, including product systems, order systems, logistics systems, payment systems, and terminal device systems. Currently, cross-border e-commerce platforms usually aggregate user behavior information through data collection and integration, and build user profile models based on user browsing history, purchase history, and transaction history. This enables applications such as user interest analysis, precise recommendations, and marketing decisions. However, due to differences in data structures and identification methods between different business systems, existing technologies often rely on simple data association methods or statistical analysis methods when integrating user behavior data, making it difficult to systematically express user behavior relationships within a unified structure.

[0003] Under the aforementioned existing technological conditions, due to the dispersed sources of user behavior data and the lack of a unified coding system, the relationships between different types of behavioral objects are difficult to express in a unified spatial structure, resulting in fragmented user behavior relationships and unclear user behavior trajectory structures. At the same time, existing user profile construction methods mostly rely on single statistical features or simple relationship networks for analysis, which are difficult to reflect the structural relationship features between user behaviors, thereby affecting the accuracy of user behavior pattern recognition and the expressive power of user profile data. Summary of the Invention

[0004] One objective of this invention is to propose a data processing method for cross-border e-commerce user profiles based on Internet of Things (IoT) coding. This invention employs an IoT coding system and a two-layer spatial folding processing method to accurately construct cross-border e-commerce user profiles, possessing the advantages of strong data integration, high behavioral correlation, and high profile accuracy.

[0005] A method for processing cross-border e-commerce user profile data based on Internet of Things (IoT) coding, according to an embodiment of the present invention, includes the following steps: Collect user behavior data, product information data, order data, logistics data, payment data, and user terminal device data from cross-border e-commerce platforms, and preprocess them to generate multi-source behavior datasets; Based on the Internet of Things (IoT) coding system, unique IoT coding identifiers are generated for user terminal devices, product objects, order objects and logistics nodes in multi-source behavior datasets, and a set of coded objects is established. A set of user behavior relationships is constructed based on the interaction relationships between the encoded objects in the set of encoded objects, including browsing relationships, transaction relationships, delivery relationships, and payment relationships; A coding behavior relationship matrix is ​​constructed based on the user behavior relationship set, and the coding behavior relationship matrix is ​​mapped to a unified behavior trajectory space to generate an initial behavior trajectory space; Perform the first-level spatial folding process on the initial behavior trajectory space to generate a unified behavior trajectory space; Based on the user behavior distribution characteristics in the unified behavior trajectory space, behavior structure parameters are generated. Based on the behavior structure parameters, a second-level spatial folding process is performed on the unified behavior trajectory space to generate a double-layer folded behavior trajectory space structure. Extract user behavior features from the double-layered folded behavior trajectory spatial structure to generate a user behavior feature set; User profile data structure is constructed from a set of user behavior features to generate cross-border e-commerce user profile data.

[0006] Optionally, the preprocessing specifically includes data cleaning, abnormal data removal, missing data completion, time index alignment, unified data format encoding, user identifier association processing, behavior event sequence construction, and data structuring processing.

[0007] Optionally, the generation of the encoded object set specifically includes: Extract device identifier, product identifier, order identifier, and logistics node identifier fields from user terminal device data, product information data, order data, and logistics data from multi-source behavior datasets. Generate a set of device objects based on the device identifier field, a set of product objects based on the product identifier field, a set of order objects based on the order identifier field, and a set of logistics node objects based on the logistics node identifier field. Obtain the object identifiers from the device object set, product object set, order object set, and logistics node object set; perform IoT code mapping processing on the object identifiers to generate the device code set, product code set, order code set, and logistics node code set; The codes in the device code set, product code set, order code set, and logistics node code set are processed to form a unified coding structure, generating a standard Internet of Things code set; The standard IoT code set is classified according to the code type field in the standard IoT code set, generating device code subsets, product code subsets, order code subsets, and logistics node code subsets; Obtain the device code subset, product code subset, order code subset, and logistics node code subset. Based on the user identifier field and device identifier field in the order data, establish the association relationship between the device code subset and the order code subset to generate a device-order association set. Based on the product identifier field in the order data, establish the association relationship between the order code subset and the product code subset to generate an order-product association set. Based on the logistics node identifier field in the logistics data, establish the association relationship between the order code subset and the logistics node code subset to generate an order-logistics node association set. Based on the device-order association set, the order-product association set, and the order-logistics node association set, the coded object association integration process is performed to generate a coded object set.

[0008] Optionally, the generation of the user behavior relationship set specifically includes: Extract the device code, product code, order code, and logistics node code from the device code subset, product code subset, order code subset, and logistics node code subset of the coded object set. Generate a browsing behavior record set based on the access records between device codes and product codes, a transaction behavior record set based on the order records between order codes and product codes, a delivery behavior record set based on the delivery records between order codes and logistics node codes, and a payment behavior record set based on the payment records between device codes and order codes. Organize the mapping relationship between device codes and product codes in the browsing behavior record set to generate a browsing relationship set; Merge the order codes and product codes in the transaction record set to generate a transaction relationship set; The correspondence between order codes and logistics node codes in the set of delivery behavior records is collected to generate a set of delivery relationships; Match the device codes and order codes in the payment behavior record set to generate a payment relationship set; By integrating the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set, a user behavior relationship set is generated.

[0009] Optionally, the generation of the initial behavior trajectory space specifically includes: Obtain the device code, product code, order code, and logistics node code from the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set in the user behavior relationship set, and generate a code node set; Perform node numbering on the set of encoded nodes, and assign a node index to each encoded node according to the node type and the order in which the nodes appear, forming a node index sequence; The node index sequence is used to perform index replacement processing on the coded nodes in the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set. The device code, product code, order code, and logistics node code are replaced with the corresponding node index to obtain the behavior relationship index set. Map the relation index pairs in the behavior relation index set to the row and column positions corresponding to the node index sequence, write the relation tag value in the corresponding matrix position, and generate the encoded behavior relation matrix; Perform relation count processing on the encoded behavior relation matrix, accumulate the occurrence of behavior relations between node indices and write them into the corresponding positions in the matrix to obtain the relation strength matrix; Based on the relation strength matrix, spatial mapping processing is performed on each coding node in the coding node set to assign behavioral trajectory coordinates to each coding node and generate a set of behavioral trajectory coordinates. The initial behavior trajectory space is formed by combining the set of behavior trajectory coordinates with the relationship strength matrix in a spatial structure.

[0010] Optionally, the generation of the unified behavior trajectory space specifically includes: The behavior trajectory coordinate set and relationship strength matrix are obtained from the initial behavior trajectory space. The device coding node, product coding node, order coding node and logistics coding node in the behavior trajectory coordinate set are identified by node number to form a trajectory node sequence. Calculate the distance between each pair of nodes in the spatial coordinates of the trajectory node sequence, and record the calculation results at the corresponding node sequence number combination position to obtain the trajectory distance set; A weighted combination operation is performed by combining the node relationship counts in the relationship strength matrix with the node distance values ​​in the trajectory distance set, and the operation result is written into the corresponding node index position to form a weighted trajectory distance set; The spatial coordinates of nodes in the trajectory node sequence are adjusted by using a weighted trajectory distance set. The node coordinates are shifted along the direction of the associated nodes and the updated node coordinates are recorded to obtain a compressed trajectory node set. Spatial region division is performed on the node coordinates in the compressed trajectory node set. Nodes whose coordinates fall within the same spatial range are classified into regions and labeled with region numbers to obtain the folded region node set. Calculate the center coordinates of each region node in the set of folded region nodes, replace the coordinates of the corresponding nodes in the region with the center coordinates, and output the set of coordinates of the folding behavior trajectory. The set of coordinates of folded behavior trajectories and the relation strength matrix are spatially integrated to generate a unified behavior trajectory space.

[0011] Optionally, the generation of the double-layer folding behavior trajectory spatial structure specifically includes: Implement node index identification for the device code nodes, product code nodes, order code nodes, and logistics code nodes in the folded behavior trajectory coordinate set in the unified behavior trajectory space to form a behavior node sequence; The number of node relationships corresponding to each node in the relationship strength matrix in the statistical behavior node sequence is recorded in the corresponding position of the node index to obtain the node relationship statistics set. Calculate the total number of relationships for each node in the node relationship statistics set, and write the calculation result to the corresponding position of the node index to form a node behavior density set; Compare the density values ​​of each node in the node behavior density set, record the difference between node densities, and obtain the behavior structure gradient set. Arrange the gradient values ​​in the set of behavioral structure gradients, and write the sorted node indices into the corresponding positions of the node numbers to generate a sorted sequence of behavioral structures. Combine the node indices, node behavior density values, and node spatial coordinates in the sorted sequence of the behavior structure, record the combination result at the corresponding position of the node index, and generate a set of behavior structure parameters. Fill the node parameters in the behavioral structure parameter set to the corresponding positions in the matrix row and column to form the behavioral structure parameter matrix; Update the node spatial coordinates in the behavior structure parameter matrix, write the updated node coordinates to the corresponding positions of the node indices, and generate a double-layer folded behavior trajectory spatial structure.

[0012] Optionally, the generation of the user behavior feature set specifically includes: Node indexing is performed on the device coding nodes, product coding nodes, order coding nodes, and logistics coding nodes in the folding behavior trajectory coordinate set in the double-layer folding behavior trajectory spatial structure to generate a behavior feature node sequence; The number of node relationships between each node in the behavior structure parameter matrix in the sequence of behavioral feature nodes is counted, and the statistical results are recorded at the corresponding positions of the node indices to generate a statistical set of node relationships. Summarize the total number of relationships between each node in the node relationship statistics set, and write the summary results to the corresponding position of the node index to obtain the node behavior frequency set; Calculate the percentage of behavior frequency of each node in the node behavior frequency set, and record the calculation results to the corresponding node index position to generate a node activity set; Arrange the activity values ​​in the node activity set, and write the sorted node indices into the corresponding positions of the node numbers to generate a behavior activity sorting sequence; The node index, node activity value, and node spatial coordinate information in the combined behavior activity ranking sequence are recorded in the corresponding position of the node index to generate a user behavior feature set.

[0013] Optionally, the generation of the cross-border e-commerce user profile data specifically includes: Obtain the node index, node activity value, and node spatial coordinate data from the user behavior feature set; perform node type identification processing on device code nodes, product code nodes, order code nodes, and logistics code nodes to generate a user behavior feature node sequence. Count the number of times the behavior corresponding to each node in the user behavior feature node sequence occurs, and record the statistical results to the corresponding position of the node index to generate a set of user behavior frequencies; Summarize the total frequency of all nodes in the user behavior frequency set, calculate the proportion of each node's behavior frequency in the total frequency, write the calculation results to the corresponding node index position, and generate a node behavior weight set. The node weight values ​​in the combined node behavior weight set and the node spatial coordinate data in the user behavior feature node sequence are combined and the combined parameters are recorded at the corresponding positions of the node indices to generate a user profile feature parameter set. Arrange the node feature parameters in the user profile feature parameter set, and record the sorted node index order to the corresponding position of the node number to generate the user profile feature sequence; Write the node feature parameters from the user profile feature sequence to the corresponding row and column positions of the matrix to generate the user profile feature matrix; The node feature parameters in the user profile feature matrix are summarized and arranged according to the node index order to generate cross-border e-commerce user profile data.

[0014] The beneficial effects of this invention are: This invention proposes a method for processing cross-border e-commerce user profile data based on IoT coding. By introducing a unified IoT coding system during the integration of multi-source behavioral data, a unified coding identifier is established for various types of objects, such as user terminal devices, product objects, order objects, and logistics nodes. This enables data objects that were originally scattered across different business systems to be uniformly associated and organized under the same coding system, thereby solving the problems of scattered sources, inconsistent data identifiers, and difficulty in establishing effective associations between behavioral objects in traditional cross-border e-commerce user behavior data. On this basis, this invention constructs a user behavior relationship set and further generates a coded behavior relationship matrix, uniformly mapping various types of user behavior relationships to the behavior trajectory space. This allows behavioral relationships in different business systems to be expressed in the same spatial structure, making the structural relationships between user behaviors clearer and significantly improving the ability to express the association between multi-source behavioral data, providing a unified data foundation for subsequent user behavior structure analysis.

[0015] This invention introduces a spatial folding mechanism based on the initial behavior trajectory space. By performing first-layer and second-layer spatial folding processes on the behavior trajectory space, the originally scattered behavior nodes are transformed into a hierarchical, double-layered folded behavior trajectory space structure. This spatial structure processing method not only effectively compresses redundant structures in user behavior relationships but also highlights key behavior nodes with high correlation and behavior density. This allows the structural features of user behavior relationships to be expressed more intuitively and stably in the spatial structure. Furthermore, by statistically and combinatorially analyzing the node behavior features in the double-layered folded behavior trajectory space structure, a set of user behavior features with structural characteristic expression capabilities can be formed. This further constructs a user profile data structure, enabling the generated cross-border e-commerce user profile data to not only reflect the frequency characteristics of user behavior but also the structural relationship characteristics between user behaviors. This improves the completeness and accuracy of the user profile data, further enhancing the data utilization effect of cross-border e-commerce platforms in application scenarios such as user behavior analysis, precise recommendation, and user group identification. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a cross-border e-commerce user profile data processing method based on Internet of Things coding proposed in this invention; Figure 2 This is a schematic diagram of the unified behavior trajectory space structure of a cross-border e-commerce user profile data processing method based on Internet of Things coding proposed in this invention. Figure 3 This is a schematic diagram of the double-layered folded behavioral trajectory spatial structure of a cross-border e-commerce user profile data processing method based on Internet of Things coding proposed in this invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0018] refer to Figures 1-3 A method for processing cross-border e-commerce user profile data based on IoT coding includes the following steps: Collect user behavior data, product information data, order data, logistics data, payment data, and user terminal device data from cross-border e-commerce platforms, and preprocess them to generate multi-source behavior datasets; Based on the Internet of Things (IoT) coding system, unique IoT coding identifiers are generated for user terminal devices, product objects, order objects and logistics nodes in multi-source behavior datasets, and a set of coded objects is established. A set of user behavior relationships is constructed based on the interaction relationships between the encoded objects in the set of encoded objects, including browsing relationships, transaction relationships, delivery relationships, and payment relationships; A coding behavior relationship matrix is ​​constructed based on the user behavior relationship set, and the coding behavior relationship matrix is ​​mapped to a unified behavior trajectory space to generate an initial behavior trajectory space; Perform the first-level spatial folding process on the initial behavior trajectory space to generate a unified behavior trajectory space; Based on the user behavior distribution characteristics in the unified behavior trajectory space, behavior structure parameters are generated. Based on the behavior structure parameters, a second-level spatial folding process is performed on the unified behavior trajectory space to generate a double-layer folded behavior trajectory space structure. Extract user behavior features from the double-layered folded behavior trajectory spatial structure to generate a user behavior feature set; User profile data structure is constructed from a set of user behavior features to generate cross-border e-commerce user profile data.

[0019] In this embodiment, preprocessing specifically includes data cleaning, abnormal data removal, missing data completion, time index alignment, unified data format encoding, user identifier association processing, behavioral event sequence construction, and data structuring processing.

[0020] In this embodiment, the generation of the encoded object set specifically includes: Extract device identifier, product identifier, order identifier, and logistics node identifier fields from user terminal device data, product information data, order data, and logistics data from multi-source behavior datasets. Generate a set of device objects based on the device identifier field, a set of product objects based on the product identifier field, a set of order objects based on the order identifier field, and a set of logistics node objects based on the logistics node identifier field. Obtain the object identifiers from the device object set, product object set, order object set, and logistics node object set; perform IoT code mapping processing on the object identifiers to generate the device code set, product code set, order code set, and logistics node code set; The codes in the device code set, product code set, order code set, and logistics node code set are processed to form a unified coding structure, generating a standard Internet of Things code set; The standard IoT code set is classified according to the code type field in the standard IoT code set, generating device code subsets, product code subsets, order code subsets, and logistics node code subsets; Obtain the device code subset, product code subset, order code subset, and logistics node code subset. Based on the user identifier field and device identifier field in the order data, establish the association relationship between the device code subset and the order code subset to generate a device-order association set. Based on the product identifier field in the order data, establish the association relationship between the order code subset and the product code subset to generate an order-product association set. Based on the logistics node identifier field in the logistics data, establish the association relationship between the order code subset and the logistics node code subset to generate an order-logistics node association set. Based on the device-order association set, the order-product association set, and the order-logistics node association set, the coded object association integration process is performed to generate a coded object set.

[0021] In this embodiment, the generation of the user behavior relationship set specifically includes: Extract the device code, product code, order code, and logistics node code from the device code subset, product code subset, order code subset, and logistics node code subset of the coded object set. Generate a browsing behavior record set based on the access records between device codes and product codes, a transaction behavior record set based on the order records between order codes and product codes, a delivery behavior record set based on the delivery records between order codes and logistics node codes, and a payment behavior record set based on the payment records between device codes and order codes. Organize the mapping relationship between device codes and product codes in the browsing behavior record set to generate a browsing relationship set; Merge the order codes and product codes in the transaction record set to generate a transaction relationship set; The correspondence between order codes and logistics node codes in the set of delivery behavior records is collected to generate a set of delivery relationships; Match the device codes and order codes in the payment behavior record set to generate a payment relationship set; By integrating the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set, a user behavior relationship set is generated.

[0022] In this embodiment, the generation of the initial behavior trajectory space specifically includes: Obtain the device code, product code, order code, and logistics node code from the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set in the user behavior relationship set, and generate a code node set; Perform node numbering on the set of encoded nodes, and assign a node index to each encoded node according to the node type and the order in which the nodes appear, forming a node index sequence; The node index sequence is used to perform index replacement processing on the coded nodes in the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set. The device code, product code, order code, and logistics node code are replaced with the corresponding node index to obtain the behavior relationship index set. Map the relation index pairs in the behavior relation index set to the row and column positions corresponding to the node index sequence, write the relation tag value in the corresponding matrix position, and generate the encoded behavior relation matrix; Perform relation count processing on the encoded behavior relation matrix, accumulate the occurrence of behavior relations between node indices and write them into the corresponding positions in the matrix to obtain the relation strength matrix; Based on the relation strength matrix, spatial mapping processing is performed on each coding node in the coding node set to assign behavioral trajectory coordinates to each coding node and generate a set of behavioral trajectory coordinates. The generation of the behavior trajectory coordinate set specifically includes: The process involves counting the number of node relationships between nodes in the relationship strength matrix, establishing a correspondence between each coded node in the coded node set and the node index in the relationship strength matrix, and generating a node relationship data set. The process also includes summarizing the number of node relationships for each coded node in the node relationship data set and writing the summarization result to the corresponding position of the node index, thus generating a total node relationship set. Based on the total node relationship set, the spatial position of each coded node in the coded node set is calculated, and spatial coordinates are determined for each coded node, generating a node spatial coordinate set. Finally, the node spatial coordinates in the node spatial coordinate set are written to the corresponding node index positions according to the node index order, generating a behavioral trajectory coordinate set. The set of behavioral trajectory coordinates and the relationship strength matrix are spatially combined to form the initial behavioral trajectory space; The generation of the initial behavior trajectory space specifically includes: Establish the correspondence between the spatial coordinates of nodes in the behavioral trajectory coordinate set and the node indices in the relationship strength matrix to generate a set of node spatial indices; write the spatial coordinates of nodes in the set of node spatial indices into the spatial node sequence positions according to the node index order to generate a spatial node sequence; establish node connection relationships at the corresponding node index positions in the spatial node sequence based on the node relationship frequency in the relationship strength matrix to generate a set of spatial relationships; combine the node connection relationship data of the spatial node sequence and the spatial relationship set to generate the initial behavioral trajectory space.

[0023] In this embodiment, the generation of the unified behavior trajectory space specifically includes: The behavior trajectory coordinate set and relationship strength matrix are obtained from the initial behavior trajectory space. The device coding node, product coding node, order coding node and logistics coding node in the behavior trajectory coordinate set are identified by node number to form a trajectory node sequence. Calculate the distance between each pair of nodes in the spatial coordinates of the trajectory node sequence, and record the calculation results at the corresponding node sequence number combination position to obtain the trajectory distance set; A weighted combination operation is performed by combining the node relationship counts in the relationship strength matrix with the node distance values ​​in the trajectory distance set, and the operation result is written into the corresponding node index position to form a weighted trajectory distance set; The spatial coordinates of nodes in the trajectory node sequence are adjusted by using a weighted trajectory distance set. The node coordinates are shifted along the direction of the associated nodes and the updated node coordinates are recorded to obtain a compressed trajectory node set. Spatial region division is performed on the node coordinates in the compressed trajectory node set. Nodes whose coordinates fall within the same spatial range are classified into regions and labeled with region numbers to obtain the folded region node set. Calculate the center coordinates of each region node in the set of folded region nodes, replace the coordinates of the corresponding nodes in the region with the center coordinates, and output the set of coordinates of the folding behavior trajectory. The set of folded behavior trajectory coordinates and the relationship strength matrix are spatially integrated to generate a unified behavior trajectory space. The generation of the unified behavior trajectory space specifically includes: Establish a correspondence between the spatial coordinates of nodes in the folded behavior trajectory coordinate set and the node indices in the relationship strength matrix. Arrange the spatial coordinates of nodes according to the node index order to generate a spatial node sequence. Based on the node relationship frequency in the relationship strength matrix, establish node connection relationships at the corresponding node index positions in the spatial node sequence and record the node connection relationships at the corresponding node index combination positions to generate a spatial relationship set. Perform spatial structure mapping between the spatial coordinates of nodes in the spatial node sequence and the node connection relationships in the spatial relationship set, and write the node spatial coordinates and node connection relationships into the corresponding node index positions to generate a trajectory structure data set. Perform structural integration of the node spatial coordinates and node connection relationships in the trajectory structure data set according to the node index order, and record the integration result into the corresponding node index positions to generate a unified behavior trajectory space.

[0024] In this embodiment, the generation of the spatial structure of the double-layer folding behavior trajectory specifically includes: Implement node index identification for the device code nodes, product code nodes, order code nodes, and logistics code nodes in the folded behavior trajectory coordinate set in the unified behavior trajectory space to form a behavior node sequence; The number of node relationships corresponding to each node in the relationship strength matrix in the statistical behavior node sequence is recorded in the corresponding position of the node index to obtain the node relationship statistics set. Calculate the total number of relationships for each node in the node relationship statistics set, and write the calculation result to the corresponding position of the node index to form a node behavior density set; Compare the density values ​​of each node in the node behavior density set, record the difference between node densities, and obtain the behavior structure gradient set. Arrange the gradient values ​​in the set of behavioral structure gradients, and write the sorted node indices into the corresponding positions of the node numbers to generate a sorted sequence of behavioral structures. Combine the node indices, node behavior density values, and node spatial coordinates in the sorted sequence of the behavior structure, record the combination result at the corresponding position of the node index, and generate a set of behavior structure parameters. Fill the node parameters in the behavioral structure parameter set to the corresponding positions in the matrix row and column to form the behavioral structure parameter matrix; Update the node spatial coordinates in the behavior structure parameter matrix, write the updated node coordinates to the corresponding positions of the node indices, and generate a double-layer folded behavior trajectory spatial structure. The generation of the spatial structure of the double-layer folding behavior trajectory specifically includes: Extract node indices, node behavior density values, and node spatial coordinates from the behavior structure parameter matrix. Arrange the node spatial coordinates according to the node index order to generate a behavior structure node sequence. Count the number of node relationships between each node in the behavior structure node sequence in the relationship strength matrix and record the statistical results in the corresponding node index combination position to generate a node relationship statistical set. Calculate the total number of node relationships for each node in the node relationship statistical set and write the calculation result to the corresponding node index position to generate a total node relationship set. Combine the total number of node relationships in the total node relationship set with the node spatial coordinates in the behavior structure node sequence to perform a spatial displacement measurement. Calculate and write the calculated node spatial coordinates to the corresponding node index position to generate the first layer of folded trajectory node set; perform spatial region division operation on the node spatial coordinates in the first layer of folded trajectory node set, merge the nodes whose spatial coordinates fall into the same spatial region and record the region number to generate the second layer of folded region node set; perform region center coordinate calculation on each region node in the second layer of folded region node set, and replace the spatial coordinates of the nodes in the region with the region center coordinates to generate the double-layer folded behavior trajectory coordinate set; write the double-layer folded behavior trajectory coordinate set to the corresponding position of the node index according to the node index order to generate the double-layer folded behavior trajectory spatial structure.

[0025] In this embodiment, the generation of the user behavior feature set specifically includes: Node indexing is performed on the device coding nodes, product coding nodes, order coding nodes, and logistics coding nodes in the folding behavior trajectory coordinate set in the double-layer folding behavior trajectory spatial structure to generate a behavior feature node sequence; The number of node relationships between each node in the behavior structure parameter matrix in the sequence of behavioral feature nodes is counted, and the statistical results are recorded at the corresponding positions of the node indices to generate a statistical set of node relationships. Summarize the total number of relationships between each node in the node relationship statistics set, and write the summary results to the corresponding position of the node index to obtain the node behavior frequency set; Calculate the percentage of behavior frequency of each node in the node behavior frequency set, and record the calculation results to the corresponding node index position to generate a node activity set; Arrange the activity values ​​in the node activity set, and write the sorted node indices into the corresponding positions of the node numbers to generate a behavior activity sorting sequence; The node index, node activity value, and node spatial coordinate information in the combined behavior activity ranking sequence are recorded in the corresponding position of the node index to generate a user behavior feature set.

[0026] In this embodiment, the generation of cross-border e-commerce user profile data specifically includes: Obtain the node index, node activity value, and node spatial coordinate data from the user behavior feature set; perform node type identification processing on device code nodes, product code nodes, order code nodes, and logistics code nodes to generate a user behavior feature node sequence. Count the number of times the behavior corresponding to each node in the user behavior feature node sequence occurs, and record the statistical results to the corresponding position of the node index to generate a set of user behavior frequencies; Summarize the total frequency of all nodes in the user behavior frequency set, calculate the proportion of each node's behavior frequency in the total frequency, write the calculation results to the corresponding node index position, and generate a node behavior weight set. The node weight values ​​in the combined node behavior weight set and the node spatial coordinate data in the user behavior feature node sequence are combined and the combined parameters are recorded at the corresponding positions of the node indices to generate a user profile feature parameter set. Arrange the node feature parameters in the user profile feature parameter set, and record the sorted node index order to the corresponding position of the node number to generate the user profile feature sequence; Write the node feature parameters from the user profile feature sequence to the corresponding row and column positions of the matrix to generate the user profile feature matrix; The node feature parameters in the user profile feature matrix are summarized and arranged according to the node index order to generate cross-border e-commerce user profile data.

[0027] Example 1: To verify the feasibility of this invention in practice, it was applied to a user behavior data processing scenario of a cross-border e-commerce platform. This platform provides global users with services such as product browsing, product purchase, cross-border logistics and delivery, and online payment. During its continuous operation, the platform has gradually accumulated a large amount of user behavior data. This data comes from multiple business data sources, including user terminal device access records, product information data, order transaction data, logistics node data, and payment behavior data. Due to differences in data structure, data identification methods, and data storage methods among different business data sources, user behavior information exhibits characteristics of scattered storage and inconsistent structure at the data level. For example, when a user browses... Access records generated when browsing products typically consist of terminal device identifiers and product identifiers. After an order transaction is completed, a record linking the order identifier and the product identifier is generated. Subsequently, during the logistics and delivery process, a delivery record linking the order identifier and the logistics node is generated. At the same time, a payment record linking the device identifier and the order identifier is generated during the payment process. Since these behavioral data exist in different data sources, traditional data processing methods often only integrate them through simple data field association or behavioral statistics. Therefore, it is difficult to fully describe the relationship between user behaviors in a unified structure, resulting in a relatively scattered user behavior trajectory structure and limited user behavior pattern recognition capabilities, thus affecting the accuracy of user profile data.

[0028] In the aforementioned application scenario, the first step is to collect and process multi-source behavioral data generated during platform operation. This data includes user terminal device access behavior data, product information data, order transaction data, logistics node data, and payment behavior data. Subsequently, the collected data undergoes data cleaning, anomaly removal, missing data completion, and unified data structure processing. This allows behavioral data from different sources to be organized within a unified data structure, forming a multi-source behavioral dataset. After the multi-source behavioral dataset is formed, an IoT coding method is introduced to generate unified coded identifiers for user terminal device objects, product objects, order objects, and logistics node objects. This enables data objects from different sources to be represented under a unified coding system. Through this coding method, behavioral objects from different data sources are mapped to unified coded objects, further forming a set of coded objects. By establishing a unified set of coded objects, data objects from different sources can be associated within the same structural space, thus providing a unified data foundation for subsequent behavioral relationship analysis.

[0029] After constructing the set of coded objects, the user's behavioral relationships on the platform are further organized based on the interaction relationships between objects in the set. Browsing behavior relationships can be obtained by analyzing access records between device codes and product codes; transaction behavior relationships can be obtained by analyzing order records between order codes and product codes; delivery behavior relationships can be obtained by analyzing delivery records between order codes and logistics node codes; and payment behavior relationships can be obtained by analyzing payment records between device codes and order codes. By aggregating and organizing these various behavioral relationships, a unified set of user behavioral relationships is formed. Based on this, the coded nodes in the user behavioral relationship set are mapped into a matrix structure. A coded behavioral relationship matrix is ​​constructed through node index replacement and relationship tag writing operations. A relationship strength matrix is ​​obtained by counting the number of relationships between nodes. After the relationship strength matrix is ​​formed, the coded nodes are mapped to the behavioral trajectory coordinate space, so that each coded node corresponds to a behavioral trajectory coordinate position in the spatial structure, thus forming a behavioral trajectory coordinate set. An initial behavioral trajectory space is further constructed. In this spatial structure, different types of behavioral nodes are expressed through spatial coordinates, and the behavioral relationships between nodes are recorded through the relationship strength matrix, thus forming a behavioral trajectory space with structural relationships.

[0030] After the initial behavior trajectory space is constructed, the node structure in the behavior trajectory space undergoes spatial folding. By comprehensively calculating the spatial distance and relationship strength between nodes, the spatial positions of nodes with high correlation are adjusted, causing related nodes to gradually aggregate in the spatial structure and further form a compressed trajectory node set. Subsequently, the nodes are divided into spatial regions based on the spatial range of their coordinates, so that nodes falling into the same region form a folded region node set. Then, the region center coordinates are calculated for each node within the region, and the original node coordinates are replaced with the region center coordinates, thus forming a folded behavior trajectory coordinate set. Through the above processing... Originally scattered behavioral nodes in space are compressed into a spatial structure with higher structural correlation, thus forming a unified behavioral trajectory space. After the unified behavioral trajectory space is formed, the structural analysis of node behavioral relationships is further carried out. By counting the number of node behavioral relationships, calculating node behavioral density, and comparing differences in node density, behavioral structure parameters are formed. Based on the behavioral structure parameters, the unified behavioral trajectory space is folded again to form a double-layer folded behavioral trajectory space structure. In this spatial structure, different user behavioral nodes are arranged according to behavioral density and behavioral structure gradient, so that the structural relationship between user behaviors can be more clearly represented in space.

[0031] After the formation of the double-layered folded behavior trajectory spatial structure, the node behavior information in the spatial structure is further analyzed. By statistically analyzing the number of relationships between nodes in the behavior structure parameter matrix, the frequency of node behavior is calculated, and the proportion of behavior frequency is further calculated, thus forming a node activity set. Subsequently, based on the node activity ranking results and node spatial coordinate information, a combination analysis is performed to generate a user behavior feature set. After the user behavior feature set is formed, parameters such as node behavior weight, node spatial coordinates, and node behavior frequency are combined and processed, and written into the matrix structure according to the node index order, thereby generating a user profile feature matrix. By summarizing and organizing the node feature parameters in the user profile feature matrix, cross-border e-commerce user profile data is finally formed.

[0032] After running continuously in this application scenario for a period of time, statistical analysis was conducted on the user behavior data processing in the platform. The statistical results show that by introducing IoT coding methods and a behavior trajectory spatial folding structure, behavioral objects from different business data sources can be associated and expressed in a unified structure. User behavior information that was originally scattered in different data sources can form a complete behavior trajectory in a unified spatial structure. At the same time, through the processing of a double-layer spatial folding structure, the relationship between behavior nodes is more clearly expressed in space, making the user behavior structure more concentrated, thereby making the user behavior feature extraction process more stable. During the continuous operation of the platform, relevant data processing records show that when processing user behavior data within the same time range, the method of the present invention can complete the integration of multi-source behavior data and behavior relationship analysis in a unified data structure, and stably output user profile data in the continuous operation environment of the platform, providing a reliable data foundation for the platform to conduct user behavior analysis and user group segmentation, thus verifying the feasibility and practicality of the method of the present invention in the cross-border e-commerce user behavior data processing scenario.

[0033] Table 1. Comparison of Overall Performance of Cross-Border E-commerce User Profile Data Processing Methods

[0034] As shown in Table 1, the method of this invention demonstrates significant advantages in several key indicators during the processing of user profile data in cross-border e-commerce. Firstly, in terms of the accuracy of user profile generation, the method of this invention achieves 89.6%, which is significantly higher than the 82.4% of traditional behavior statistics methods and the 85.7% of traditional graph embedding methods. This indicates that by introducing an IoT coding system and a unified behavior trajectory spatial structure, user behaviors from different data sources can be expressed in a unified structure, thereby reducing the behavior recognition bias caused by data dispersion in traditional methods. At the same time, in terms of user behavior association recognition rate, the method of this invention achieves 87.9%, which is significantly higher than the 79.2% of traditional behavior statistics methods and the 83.6% of traditional graph embedding methods. This shows that through behavior trajectory spatial modeling and spatial folding structure processing, the association between user behaviors can be more completely identified.

[0035] In terms of data processing efficiency, the data processing throughput of the method of this invention reaches 248 data entries / second, while the traditional behavior statistics method is 163 data entries / second and the traditional graph embedding method is 191 data entries / second. This indicator shows that the method of this invention has higher processing efficiency in the process of multi-source data integration and behavior relationship processing. This is mainly because, under the Internet of Things coding system, data objects from different sources can directly establish association relationships through unified coding, thereby reducing the calculation process of multiple data matching and field association required in traditional methods, thus improving the overall processing efficiency. At the same time, in terms of behavior trajectory space construction time, the method of this invention only requires 57 seconds, while the traditional behavior statistics method requires 94 seconds and the traditional graph embedding method requires 79 seconds. This shows that after compressing and aggregating behavior nodes through the spatial folding structure, the behavior trajectory space construction process can be completed in a shorter time.

[0036] Further analysis of the user profile update cycle index shows that the method of this invention requires approximately 21 minutes to update the user profile, while the traditional behavior statistics method takes 34 minutes and the traditional graph embedding method takes 29 minutes. This result indicates that in a continuously running cross-border e-commerce platform environment, the method of this invention can complete the update of user behavior data and profile generation faster, enabling the user profile to reflect changes in user behavior more promptly. Furthermore, in terms of data memory usage, the method of this invention uses 468MB, significantly lower than the 612MB of the traditional behavior statistics method and the 556MB of the traditional graph embedding method. This demonstrates that by compressing the behavior trajectory space through a double-layer spatial folding structure, redundant structures between behavior nodes can be reduced, thereby lowering the overall data storage and computing resource consumption.

[0037] As can be seen from the data in Table 1, the method of the present invention outperforms traditional methods in terms of user profile generation accuracy, behavior association recognition capability, data processing efficiency, and resource utilization. The main reason for its performance improvement is that the present invention uses an IoT coding system to uniformly identify multi-source behavioral data, enabling data objects from different sources to be associated and expressed in a unified structure. At the same time, by constructing a behavior trajectory space and implementing double-layer spatial folding processing, the structural relationship between user behaviors is compressed and strengthened in space, thereby enabling more accurate extraction of user behavior features and generation of user profile data. The above results show that the method of the present invention has good practical application effects and stable data processing capabilities in the process of cross-border e-commerce user profile data processing.

[0038] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for processing cross-border e-commerce user profile data based on Internet of Things (IoT) coding, characterized in that, Includes the following steps: Collect user behavior data, product information data, order data, logistics data, payment data, and user terminal device data from cross-border e-commerce platforms, and preprocess them to generate multi-source behavior datasets; Based on the Internet of Things (IoT) coding system, unique IoT coding identifiers are generated for user terminal devices, product objects, order objects and logistics nodes in multi-source behavior datasets, and a set of coded objects is established. A set of user behavior relationships is constructed based on the interaction relationships between the encoded objects in the set of encoded objects, including browsing relationships, transaction relationships, delivery relationships, and payment relationships; A coding behavior relationship matrix is ​​constructed based on the user behavior relationship set, and the coding behavior relationship matrix is ​​mapped to a unified behavior trajectory space to generate an initial behavior trajectory space; Perform the first-level spatial folding process on the initial behavior trajectory space to generate a unified behavior trajectory space; Based on the user behavior distribution characteristics in the unified behavior trajectory space, behavior structure parameters are generated. Based on the behavior structure parameters, a second-level spatial folding process is performed on the unified behavior trajectory space to generate a double-layer folded behavior trajectory space structure. Extract user behavior features from the double-layered folded behavior trajectory spatial structure to generate a set of user behavior features; User profile data structure is constructed from a set of user behavior features to generate cross-border e-commerce user profile data.

2. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The preprocessing specifically includes data cleaning, abnormal data removal, missing data completion, time index alignment, unified data format encoding, user identifier association processing, behavioral event sequence construction, and data structuring.

3. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the encoded object set specifically includes: Extract device identifier, product identifier, order identifier, and logistics node identifier fields from user terminal device data, product information data, order data, and logistics data from multi-source behavior datasets. Generate a set of device objects based on the device identifier field, a set of product objects based on the product identifier field, a set of order objects based on the order identifier field, and a set of logistics node objects based on the logistics node identifier field. Obtain the object identifiers from the device object set, product object set, order object set, and logistics node object set; perform IoT code mapping processing on the object identifiers to generate the device code set, product code set, order code set, and logistics node code set; The codes in the device code set, product code set, order code set, and logistics node code set are processed to form a unified coding structure, generating a standard Internet of Things code set; The standard IoT code set is classified according to the code type field in the standard IoT code set, generating device code subsets, product code subsets, order code subsets, and logistics node code subsets; Obtain the device code subset, product code subset, order code subset, and logistics node code subset. Based on the user identifier field and device identifier field in the order data, establish the association relationship between the device code subset and the order code subset to generate a device-order association set. Based on the product identifier field in the order data, establish the association relationship between the order code subset and the product code subset to generate an order-product association set. Based on the logistics node identifier field in the logistics data, establish the association relationship between the order code subset and the logistics node code subset to generate an order-logistics node association set. Based on the device-order association set, the order-product association set, and the order-logistics node association set, the coded object association integration process is performed to generate a coded object set.

4. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the user behavior relationship set specifically includes: Extract the device code, product code, order code, and logistics node code from the device code subset, product code subset, order code subset, and logistics node code subset of the coded object set. Generate a browsing behavior record set based on the access records between device codes and product codes, a transaction behavior record set based on the order records between order codes and product codes, a delivery behavior record set based on the delivery records between order codes and logistics node codes, and a payment behavior record set based on the payment records between device codes and order codes. Organize the mapping relationship between device codes and product codes in the browsing behavior record set to generate a browsing relationship set; Merge the order codes and product codes in the transaction record set to generate a transaction relationship set; The correspondence between order codes and logistics node codes in the set of delivery behavior records is collected to generate a set of delivery relationships; Match the device codes and order codes in the payment behavior record set to generate a payment relationship set; By integrating the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set, a user behavior relationship set is generated.

5. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the initial behavior trajectory space specifically includes: Obtain device codes, product codes, order codes, and logistics node codes from the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set in the user behavior relationship set, and generate a set of coded nodes; Perform node numbering on the set of encoded nodes, and assign a node index to each encoded node according to the node type and the order in which the nodes appear, forming a node index sequence; The node index sequence is used to perform index replacement processing on the coded nodes in the browsing relationship set, transaction relationship set, delivery relationship set, and payment relationship set. The device code, product code, order code, and logistics node code are replaced with the corresponding node index to obtain the behavior relationship index set. Map the relation index pairs in the behavior relation index set to the row and column positions corresponding to the node index sequence, write the relation tag value in the corresponding matrix position, and generate the encoded behavior relation matrix; Perform relation count processing on the encoded behavior relation matrix, accumulate the occurrence of behavior relations between node indices and write them into the corresponding positions in the matrix to obtain the relation strength matrix; Based on the relation strength matrix, spatial mapping processing is performed on each coding node in the coding node set to assign behavioral trajectory coordinates to each coding node and generate a set of behavioral trajectory coordinates. The initial behavior trajectory space is formed by combining the set of behavior trajectory coordinates with the relationship strength matrix in a spatial structure.

6. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the unified behavior trajectory space specifically includes: The behavior trajectory coordinate set and relationship strength matrix are obtained from the initial behavior trajectory space. The device coding node, product coding node, order coding node and logistics coding node in the behavior trajectory coordinate set are identified by node number to form a trajectory node sequence. Calculate the distance between each pair of nodes in the spatial coordinates of the trajectory node sequence, and record the calculation results at the corresponding node sequence number combination position to obtain the trajectory distance set; A weighted combination operation is performed by combining the node relationship counts in the relationship strength matrix with the node distance values ​​in the trajectory distance set, and the operation result is written into the corresponding node index position to form a weighted trajectory distance set; The spatial coordinates of nodes in the trajectory node sequence are adjusted by using a weighted trajectory distance set. The node coordinates are shifted along the direction of the associated nodes and the updated node coordinates are recorded to obtain a compressed trajectory node set. Spatial region division is performed on the node coordinates in the compressed trajectory node set. Nodes whose coordinates fall within the same spatial range are classified into regions and labeled with region numbers to obtain the folded region node set. Calculate the center coordinates of each region node in the set of folded region nodes, replace the coordinates of the corresponding nodes in the region with the center coordinates, and output the set of coordinates of the folding behavior trajectory. The set of coordinates of folded behavior trajectories and the relation strength matrix are spatially integrated to generate a unified behavior trajectory space.

7. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the spatial structure of the double-layer folding behavior trajectory specifically includes: Implement node index identification for the device code nodes, product code nodes, order code nodes, and logistics code nodes in the folded behavior trajectory coordinate set in the unified behavior trajectory space to form a behavior node sequence; The number of node relationships corresponding to each node in the relationship strength matrix in the statistical behavior node sequence is recorded in the corresponding position of the node index to obtain the node relationship statistics set. Calculate the total number of relationships for each node in the node relationship statistics set, and write the calculation result to the corresponding position of the node index to form a node behavior density set; Compare the density values ​​of each node in the node behavior density set, record the difference between node densities, and obtain the behavior structure gradient set. Arrange the gradient values ​​in the set of behavioral structure gradients, and write the sorted node indices into the corresponding positions of the node numbers to generate a sorted sequence of behavioral structures. Combine the node indices, node behavior density values, and node spatial coordinates in the sorted sequence of the behavior structure, record the combination result at the corresponding position of the node index, and generate a set of behavior structure parameters. Fill the node parameters in the behavioral structure parameter set to the corresponding positions in the matrix row and column to form the behavioral structure parameter matrix; Update the node spatial coordinates in the behavior structure parameter matrix, write the updated node coordinates to the corresponding positions of the node indices, and generate a double-layer folded behavior trajectory spatial structure.

8. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the user behavior feature set specifically includes: Node indexing is performed on the device coding nodes, product coding nodes, order coding nodes, and logistics coding nodes in the folding behavior trajectory coordinate set in the double-layer folding behavior trajectory spatial structure to generate a behavior feature node sequence; The number of node relationships between each node in the behavior structure parameter matrix in the sequence of behavioral feature nodes is counted, and the statistical results are recorded at the corresponding positions of the node indices to generate a statistical set of node relationships. Summarize the total number of relationships between each node in the node relationship statistics set, and write the summary results to the corresponding position of the node index to obtain the node behavior frequency set; Calculate the percentage of behavior frequency of each node in the node behavior frequency set, and record the calculation results to the corresponding node index position to generate a node activity set; Arrange the activity values ​​in the node activity set, and write the sorted node indices into the corresponding positions of the node numbers to generate a behavior activity sorting sequence; The node index, node activity value, and node spatial coordinate information in the combined behavior activity ranking sequence are recorded in the corresponding position of the node index to generate a user behavior feature set.

9. The method for processing cross-border e-commerce user profile data based on IoT coding according to claim 1, characterized in that, The generation of the cross-border e-commerce user profile data specifically includes: Obtain node indexes, node activity values, and node spatial coordinate data from the user behavior feature set; perform node type identification processing on device coding nodes, product coding nodes, order coding nodes, and logistics coding nodes to generate a user behavior feature node sequence. Count the number of times the behavior corresponding to each node in the user behavior feature node sequence occurs, and record the statistical results to the corresponding position of the node index to generate a set of user behavior frequencies; Summarize the total frequency of all nodes in the user behavior frequency set, calculate the proportion of each node's behavior frequency in the total frequency, write the calculation results to the corresponding node index position, and generate a node behavior weight set. The node weight values ​​in the combined node behavior weight set and the node spatial coordinate data in the user behavior feature node sequence are combined and the combined parameters are recorded at the corresponding positions of the node indices to generate a user profile feature parameter set. Arrange the node feature parameters in the user profile feature parameter set, and record the sorted node index order to the corresponding position of the node number to generate the user profile feature sequence; Write the node feature parameters from the user profile feature sequence to the corresponding row and column positions of the matrix to generate the user profile feature matrix; The node feature parameters in the user profile feature matrix are summarized and arranged according to the node index order to generate cross-border e-commerce user profile data.

Images