Cross-border e-commerce order management system based on presale data analysis

Abstract

The invention relates to the technical field of cross-border e-commerce, in particular to a cross-border e-commerce order management system based on presale data analysis. The system comprises a system connection module, an address processing general control module and a user interaction feedback module, a prefix tree map construction module constructs a prefix tree map model through map information obtained by the system connection module, and then a user filling address in presale order data obtained by the system connection module is input to an address structure analysis module; an address structure analysis module converts an abbreviation in a filling address into a filling address with the format the same as that of map information, then whether the address filled by a user is less filled or not is judged through a prefix tree map model, and when less filling occurs, the map information which is less filled is filled through the map information, so that the filling efficiency of the user is improved. And if the less-filled information is an end node of the prefix tree map model in the process, the user interaction feedback module analyzes the last byte and resets the last byte, so that the prefix tree map model is adjusted.

Description

Technical Field

[0001] The present invention relates to the technical field of cross-border e-commerce, and more specifically, to a cross-border e-commerce order management system based on pre-sale data analysis. Background Art

[0002] In the current cross-border e-commerce order management system, after a user makes a purchase, the user needs to fill in order data so that the goods can be delivered to the user in a timely manner. At the same time, to avoid errors in the order data filled in by the user, mandatory fields will be set in the user order placement process. For example, the mobile phone number must conform to the number format specification of the target country or region, and the email address must meet the standard email format. The cross-border e-commerce order management system checks in real time when the user enters. If the format does not match, it will immediately prompt for modification;

[0003] However, in the actual filling process, for the data corresponding to the address filled in by the user in the order data, in order to protect the user's privacy, the user may deliberately not fill in the complete or detailed address information and only provide a general range. Therefore, the filled address cannot be further restricted. At this time, situations such as incorrect filling or incomplete filling by the user will occur, resulting in problems such as incorrect package delivery and delivery delays in the subsequent delivery process, which will cause unnecessary trouble to both the user and the e-commerce enterprise. In view of this, we propose a cross-border e-commerce order management system based on pre-sale data analysis. Summary of the Invention

[0004] The purpose of the present invention is to solve the problem that it is impossible to timely judge when there is missing filling or incomplete filling in the address information filled in the user order output.

[0005] To achieve the above purpose, the present invention provides a cross-border e-commerce order management system based on pre-sale data analysis, including a system connection module, an address processing general control module, and a user interaction feedback module;

[0006] The system connection module establishes connections with the cross-border e-commerce system and the map system to obtain pre-sale order data, map information, and address hierarchical structures;

[0007] The address processing general control module includes a prefix tree map construction module and an address structure analysis module;

[0008] The prefix tree map construction module perceives the map information obtained by the system connection module and constructs a prefix tree map model through the map information. Each structural point in the prefix tree map model includes character data, a list of child nodes, and end marker information. The address structure analysis module uses a fuzzy matching algorithm to convert the abbreviations that appear in the filled address into the filled address in the same format as the map information, and then inputs the converted user-filled address into the prefix tree map model to determine whether there is a situation of incomplete filling or error in the user-filled address. If not, it is determined that the user-filled address is accurate;

[0009] The user interaction feedback module is used to sense the situation where the user fills in an incomplete address, analyze whether the incomplete address is an end node in the prefix tree map model, and record the number of users with the incomplete address as the end node. And the number of users with the incomplete address as the previous node of the end node. If , the node optimization method is used to optimize the end nodes in the prefix tree map model.

[0010] As a further improvement of this technical solution, the system connection module applies for API information, request methods, required request parameters, and authentication methods from the cross-border e-commerce system and the map system respectively;

[0011] Then, the network request library in the programming language is used to construct an HTTP request, authentication information is added to the HTTP request, and the constructed request is sent to the corresponding servers of the cross-border e-commerce system and the map system through the network. The request follows the format and rules specified by the HTTP protocol in the network and is transmitted in the form of data packets.

[0012] After receiving the request, the servers of the cross-border e-commerce system and the map system parse the request. If the request is legal, the servers of the cross-border e-commerce system and the map system query the pre-sale order data and map information according to the request content. The cross-border e-commerce system and the map system servers respectively encapsulate the queried pre-sale order data and map information into response messages, return the response messages in JSON format, and the response messages also follow the HTTP protocol and are transmitted in the form of data packets in the network;

[0013] After receiving the response message, the system connection module uses the corresponding parsing library to convert the data into the data structure in the programming language according to the response message format, so as to obtain the pre-sale order data and map information.

[0014] As a further improvement of this technical solution, the prefix tree map construction module constructs a prefix tree map model, specifically through the following steps:

[0015] Step 1: Define the prefix tree node structure through the address hierarchical structure. Each node contains character data, a list of child nodes, and end marker information;

[0016] In the overall prefix tree node structure, each character data represents a character in the address, the list of child nodes stores all the child nodes of the node, and the end marker indicates whether it is the end of a complete address;

[0017] Step 2: Address splitting and insertion:

[0018] The predefined address hierarchical structure in the map system of the perception system connection module is traversed for each address information in the map. The addresses in the map information are split through the address hierarchical structure, and the address information is sorted according to the address hierarchical structure to determine the character data corresponding to each address.

[0019] Start operating from the root node of the prefix tree. Insert each character data in the address into the prefix tree in turn. During the insertion process, check the list of child nodes of the current node. If there is a node corresponding to the character data, move the character data to the child node and continue to insert the next character data according to the address hierarchy. If not, create a new node, add the character data to the list of child nodes of the current node, and then move to the newly created node.

[0020] Step 3: When traversing all the digits corresponding to the entire address, set the end flag of the end node to True, indicating that the path formed from the root node to this node represents a complete address.

[0021] As a further improvement of this technical solution, the steps of the fuzzy matching algorithm in the address structure analysis module are as follows:

[0022] Step 1: Sense the address hierarchical structures in the user-entered address and the map information respectively. Combine the addresses at the first level in the map information into the first address library, the addresses at the second level into the second address library, until all the addresses in the map information are combined into address libraries.

[0023] Retrieve the addresses corresponding to different hierarchical structures from the user-entered address, and define them as the first address, the second address, until all the addresses in the entered address are defined.

[0024] Step 2: Match the first address with the first-level address library. If there is the same address, it is determined that the first-level address entered by the user is correct. Otherwise, calculate the similarity between the first-level address and all the addresses in the first-level address library, and mark the address with the maximum similarity in the first-level address library.

[0025] Set a similarity threshold. If the maximum similarity > similarity threshold, replace the user-entered address with the address with the maximum similarity in the first-level address library.

[0026] Step 3: After the first address is matched with the first-level address library, retrieve the second address again and match it with the second-level address library until all the addresses corresponding to the entered address are matched.

[0027] As a further improvement of this technical solution, the calculation method for calculating the similarity between the first-level address and all the addresses in the first-level address library in the fuzzy matching algorithm of the address structure analysis module is as follows:

[0028] A method for establishing a connection with a cross-border e-commerce system through a system connection module, establishing a connection with a text corpus system, segmenting the text corpus, and assigning a row and column position to each vocabulary in a matrix;

[0029] Calculate the co-occurrence of the first-level addresses in the text corpus and all addresses in the first address library respectively. If the first-level address Y and an address O in the first address library appear in the same text, the value at the corresponding position in the co-occurrence matrix will increase, and count the vocabulary that appears together with the first-level address and all addresses in the first address library, and construct a co-occurrence matrix that records the co-occurrence times of the vocabulary;

[0030] After obtaining the co-occurrence matrix, the first-level address and all addresses in the first address library each correspond to a row vector or column vector in the matrix. The vector corresponding to the first-level address Y is and an address O in the first address library is . At this time, calculate the similarity between the vectors by the cosine similarity calculation method.

[0031] As a further improvement of this technical solution, the address structure analysis module inputs the filled address converted into the same format as the map information into the prefix tree map model constructed by the prefix tree map construction module, splits the filled address according to the address hierarchical structure, and determines the corresponding character data , and inserts the character data into the prefix tree map model;

[0032] Perceive that the prefix tree map model is , where is a list of different child nodes in the prefix tree map model, and there are different child nodes in the list of child nodes, such as , , , ;

[0033] During the process of inserting into the prefix tree map model:

[0034] If there is no corresponding node for a certain character data in the filled address in the prefix tree map model, it is determined that there is missing address information in the user's filled address. At this time, skip the node with missing address information, and retrieve the sub-list and the next node list of the node where there is no corresponding node in the prefix tree map model, and match the un-matched character data in the filled address again;

[0035] For example , it is determined that there is missing address information in the user's filled address ;

[0036] And fill in the missing address information according to the address information adjacent to the missing address information in the address hierarchical structure.

[0037] As a further improvement of this technical solution, if a certain character data of the filled address does not have a corresponding node in the prefix tree map model in the address structure analysis module, it is determined that there is misfilled address information in the user-filled address, and the misfilled address information is deleted and converted into missing address information;

[0038] Such as , it is determined that there is misfilled address information in the user-filled address;

[0039] If the end node is not True after all the character data of the filled address is inserted into the prefix tree map model, but the previous nodes and character data are exactly the same, it is also determined that there is missing address information in the user-filled address; such as , it is determined that the end node corresponding to the prefix tree map model is abnormal;

[0040] If the end node is True after all the character data of the filled address is inserted into the prefix tree map model, it is determined that the user-filled address is correct, such as , it is determined that the user-filled address is correct.

[0041] As a further improvement of this technical solution, the address processing master control module further includes an address error correction and filling module. The address error correction and filling module is used to sense missing addresses and misfilled address information, fill in the missing addresses according to the address hierarchical structure, modify the misfilled address information, and establish a connection with the cross-border e-commerce system through the system connection module for the filled and modified address information, and input the corresponding missing addresses and misfilled address information after filling and repair into the cross-border e-commerce system to prompt the user to confirm the address information.

[0042] As a further improvement of this technical solution, the user interaction feedback module senses the end node of the prefix tree map model corresponding to the character data of the current user, and senses the address information in the historical order data of the cross-border e-commerce system, and converts all the historical address information into corresponding character data, which is defined as the historical character data set;

[0043] Match and calculate the current character data that is exactly the same as the historical character data set, and record the number of users ;

[0044] Exclude the end node in the historical character data set , and match and calculate the current character data that is exactly the same as the excluded end node , and record the number of users as .

[0045] As a further improvement of this technical solution, in the user interaction feedback module, when the node optimization method is used, if , then the end node of the prefix tree map model in the address processing master control module is removed, and the node before the end node is used as the new node , otherwise the end node is not optimized.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] In the cross-border e-commerce order management system based on pre-sale data analysis, the prefix tree map construction module constructs a prefix tree map model through the map information obtained by the system connection module, and then inputs the user-entered address in the pre-sale order data obtained by the system connection module into the address structure analysis module. The address structure analysis module converts the abbreviations in the filled address into the filled address in the same format as the map information, and then determines whether there is a situation of missing filling in the user-entered address through the prefix tree map model. When there is a situation of missing filling, the missing map information is filled through the map information. And in this process, if the missing information is the end node of the prefix tree map model, the user interaction feedback module analyzes and resets the last byte, so as to adjust the prefix tree map model, flexibly adjust according to the actual situation, enhance the adaptability of the system to different address filling situations, avoid being unable to handle special situations due to fixed model limitations, and at the same time further improve the accuracy of determining whether the user-entered address is correct, avoiding problems such as incorrect or missing filling of user addresses resulting in incorrect package delivery and delivery delays. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 is the overall module schematic diagram of the present invention;

[0049] Figure 2 is the schematic diagram of the address processing master control module of the present invention;

[0050] Figure 3 is the working principle flow chart of the present invention.

[0051] The meanings of the various reference numerals in the figure are as follows:

[0052] 100, system connection module; 200, address processing master control module; 210, prefix tree map construction module; 220, address structure analysis module; 230, address error correction and filling module; 300, user interaction feedback module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0054] The system connection module 100 establishes connections with the cross-border e-commerce system and the map system, obtains pre-sale order data, map information (the map information includes address data at various levels, such as detailed information of countries, cities, streets, communities, etc., as well as auxiliary data such as relevant geographical coordinates and address codes), and the address hierarchical structure.

[0055] The system connection module 100 applies for API information from the cross-border e-commerce system and the map system respectively. The API information includes the endpoints of the API, request methods (such as GET, POST, PUT, DELETE, etc.), required request parameters, and authentication methods, etc.

[0056] The system connection module 100 uses the network request library in the programming language to construct an HTTP request (the HTTP request includes pre-sale order data and map information), adds authentication information to the HTTP request, and then sends the constructed requests to the corresponding servers of the cross-border e-commerce system and the map system respectively through the network. The requests follow the format and rules specified by the HTTP protocol in the network and are transmitted in the form of data packets.

[0057] After receiving the requests, the servers of the cross-border e-commerce system and the map system parse the requests and check the legality of the requests, including whether the request format is correct, whether the parameters are complete, whether the authentication information is valid, etc. If the requests are legal, the servers of the cross-border e-commerce system and the map system query the pre-sale order data and map information according to the request content. The servers of the cross-border e-commerce system and the map system respectively encapsulate the queried pre-sale order data and map information into response messages, return the response messages in JSON format, and the response messages also follow the HTTP and are transmitted in the form of data packets in the network.

[0058] After receiving the response messages, the system connection module 100 converts the data into the data structure in the programming language using the corresponding parsing library according to the response message format, so as to obtain the pre-sale order data and map information.

[0059] In the current cross-border e-commerce order management system, after a user makes a purchase, the user needs to fill in order data so that the goods can be delivered to the user in a timely manner. At the same time, to avoid errors in the order data filled in by the user, mandatory fields will be set in the user's order placement process. For example, the mobile phone number must conform to the number format specifications of the target country or region, and the email address must meet the standard email format. The cross-border e-commerce order management system checks in real-time when the user enters. If the format does not match, it will immediately prompt for modification;

[0060] However, in the actual filling process, since the data corresponding to the address filled in by the user in the order data, in order to protect the user's privacy, the user may deliberately not fill in the complete or detailed address information and only provide a general range. Therefore, it is not possible to further limit the filled address. At this time, situations such as incorrect filling or incomplete filling by the user will occur, resulting in problems such as incorrect package delivery and delivery delays in the subsequent delivery process, which will bring unnecessary troubles to both the user and the e-commerce enterprise;

[0061] The address processing master control module 200 includes a prefix tree map construction module 210 and an address structure analysis module 220;

[0062] The prefix tree map construction module 210 constructs a prefix tree map model based on the map information obtained from the system connection module 100, specifically through the following steps:

[0063] Step 1: Define the prefix tree node structure through the address hierarchical structure. Each node contains character data, a list of child nodes, and end marker information;

[0064] Each character data in the prefix tree node structure represents a character in the address. The list of child nodes stores all the child nodes of this node. The end marker indicates whether it is the end of a complete address;

[0065] Step 2: Address splitting and insertion:

[0066] Perceive the "country-province / state-city-district-street-house number" address hierarchical structure predefined in the map system of the system connection module 100. Traverse each address information in the map, split the address in the map information through the address hierarchical structure, sort the address information according to the address hierarchical structure, and determine the character data corresponding to each address;

[0067] Start operating from the root node of the prefix tree, and insert each character data in the address into the prefix tree in sequence. During the insertion process, check the list of child nodes of the current node. If there is a node corresponding to the character data, move the character data to the child node and continue to insert the next character data according to the address hierarchy. If not, create a new node, add the character data to the list of child nodes of the current node, and then move to the newly created node;

[0068] Step 3: When all the digits corresponding to the entire address have been traversed, set the end flag of the end node to True, indicating that the path formed from the root node to this node represents a complete address.

[0069] In the digital age, all industries are pursuing data standardization and normalization to achieve more efficient information circulation and business collaboration. For cross-border e-commerce and the map industry to accurately plan delivery routes for logistics personnel, the address information filled in the existing cross-border e-commerce system and the address hierarchy structure in the map system are the same;

[0070] The address structure analysis module 220 senses the address filled in by the user in the pre-sale order data, and uses a fuzzy matching algorithm to convert the abbreviations in the filled address into the filled address in the same format as the map information, and then inputs it into the prefix tree map model constructed by the prefix tree map construction module 210. The filled address is split according to the address hierarchy structure to determine the corresponding character data , and insert the character data into the prefix tree map model;

[0071] Sense the prefix tree map model as , where is a different list of child nodes in the prefix tree map model, and there are different child nodes in the list of child nodes, such as , , , ;

[0072] During the process of inserting into the prefix tree map model:

[0073] If there is no corresponding node for a certain character data in the filled address in the prefix tree map model, it is determined that there is missing address information in the user-filled address. At this time, skip the node with missing address information, and retrieve the sub-list and the next node list where the corresponding node does not exist in the prefix tree map model, and match the unmatched character data in the filled address again;

[0074] For example , it is determined that there is missing address information in the user-filled address ;

[0075] The address structure analysis module 220 fills in the missing address information based on the address information adjacent to the missing address information in the address hierarchical structure. For example, if the user has filled in "China - Guangdong Province - Guangzhou City - No. x", lacking the "district" and "street" information, and there are different components of addresses in the address hierarchical structure, at this time, according to the positional relationship between "Guangzhou City" and "No. x", the "x district" and "x street" information where "No. x" is located is obtained, and then the information is filled into the address filled in by the user;

[0076] If a certain character data of the filled address does not have a corresponding node in the prefix tree map model, it is determined that there is misfilled address information in the user - filled address, the misfilled address information is deleted, and it is converted into missing address information;

[0077] Such as , it is determined that there is misfilled address information in the user - filled address;

[0078] If the end node is not True after all the character data of the filled address is inserted into the prefix tree map model, but the nodes and character data before that are exactly the same, it is also determined that there is missing address information in the user - filled address; Such as , it is determined that the end node corresponding to the prefix tree map model is abnormal;

[0079] If the end node is True after all the character data of the filled address is inserted into the prefix tree map model, it is determined that the user - filled address is correct. Such as , it is determined that the user - filled address is correct.

[0080] In order to avoid the situation where when the prefix tree map model determines whether the user - filled address is correct, due to differences in the filling habits of different users, different address information is abbreviated (such as writing "Beijing City" as "Beijing", "Xihu District" as "Xihu", etc.). At this time, if the abbreviated address information cannot be modified, it will affect the determination of the prefix tree map model.

[0081] The working steps of the fuzzy matching algorithm in the address structure analysis module 220 to convert the abbreviations that appear in the filled address into the filled address in the same format as the map information are as follows:

[0082] Step 1: Sense the address hierarchical structures in the user - filled address and the map information respectively. Combine the addresses at the first level in the map information into the first address library, the addresses at the second level into the second address library, until all the addresses in the map information are combined into address libraries;

[0083] Retrieve the addresses corresponding to different hierarchical structures from the user - filled address, and define them as the first address, the second address, until all the addresses in the filled address are defined;

[0084] Step 2: Match the first address with the first-level address library. If there is a same address, it is determined that the first-level address filled in by the user is correct; otherwise, calculate the similarity between the first-level address and all addresses in the first-level address library, and mark the address with the maximum similarity in the first-level address library.

[0085] Set a similarity threshold. If the maximum similarity > the similarity threshold, replace the address filled in by the user with the address with the maximum similarity in the first-level address library.

[0086] Step 3: After the matching between the first address and the first-level address library is completed, call out the second address again to match with the second-level address library until all the addresses corresponding to the filled-in addresses are matched.

[0087] Calculate the similarity between the first-level address and all addresses in the first-level address library. By the method of establishing a connection with the cross-border e-commerce system through the system connection module 100, establish a connection with the text corpus system (the text corpus can be a set of documents, a large database, or a text collection in a specific field), segment the text corpus, and assign a row and column position to each vocabulary in the matrix.

[0088] Calculate the co-occurrence situation of the first-level address in the text corpus and all addresses in the first address library respectively. If the first-level address Y and a certain address O in the first address library appear in the same text, the value at the corresponding position (row Y and column O, and row O and column Y) in the co-occurrence matrix will increase by 1, and count the vocabulary that the first-level address and all addresses in the first address library appear together, and construct a co-occurrence matrix that records the co-occurrence times of the vocabulary.

[0089] After obtaining the co-occurrence matrix, the first-level address and all addresses in the first address library each correspond to a row vector or column vector in the matrix. The vector corresponding to the first-level address Y is and a certain address O in the first address library is , at this time, calculate the similarity between the vectors by the cosine similarity calculation method, and the calculation formula is as follows:

[0090] ;

[0091] The value range of the cosine similarity is. The closer the value is to, the smaller the included angle between the two vectors, and the higher the similarity of the vocabulary; the closer the value is to, the larger the included angle between the two vectors, and the greater the difference of the vocabulary; when the value is, the two vectors are orthogonal, that is, there is no similarity.

[0092] The address processing master control module 200 further includes an address error correction and filling module 230. The address error correction and filling module 230 is used to detect missing or incorrect address information, fill in the missing addresses and modify the incorrect address information according to the address hierarchical structure, and input the filled and modified address information into the cross-border e-commerce system through the connection established by the system connection module 100 with the cross-border e-commerce system, and prompt the user to confirm the address information.

[0093] The present invention also takes into account that in the existing map system, the map information is too detailed. In the cross-border e-commerce system, due to concerns about address information leakage, users often only describe the address roughly when filling in the address to fit their actual address. Therefore, the prefix tree map model established by the prefix tree map construction module 210 through the map information in the map system will be too detailed. If the prefix tree map model in the prefix tree map construction module 210 cannot be adjusted in time, it will be interfered with when determining whether the user's filled address is correct, affecting the accuracy of the judgment.

[0094] The user interaction feedback module 300 senses the end node of the prefix tree map model corresponding to the character data of the current user, and senses the address information in the historical order data of the cross-border e-commerce system, converts all the historical address information into corresponding character data, and defines it as the historical character data set;

[0095] Match and calculate the current character data that is exactly the same as the historical character data set, and record the number of users ;

[0096] Delete the end node in the historical character data set and match and calculate the current character data that is exactly the same as the deleted end node again The number of users recorded is ;

[0097] If , then delete the end node of the prefix tree map model in the address processing master control module 200 and use the previous node of the end node as the new node , otherwise do not optimize the end node .

[0098] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A cross-border e-commerce order management system based on pre-sale data analysis, characterized by: It comprises a system connection module (100), an address processing master control module (200) and a user interaction feedback module (300); The system connection module (100) establishes a connection with the cross-border e-commerce system and the map system to obtain pre-sale order data, map information and address hierarchical structure; The address processing master control module (200) comprises a prefix tree map construction module (210) and an address structure analysis module (220), wherein the prefix tree map construction module (210) constructs a prefix tree map model through map information, character data, a subnode list and end mark information in each structure point structure in the prefix tree map model, and the address structure analysis module (220) uses a fuzzy matching algorithm to convert the abbreviation appearing in the filled-in address into a filled-in address in the same format as the map information, and inputs the abbreviation into the prefix tree map model to determine whether there is any missing or wrong address filled in by the user, and if not, it is determined that the address filled in by the user is correct; The user interaction feedback module (300) is used to sense the situation where the user has underfilled the address, analyze whether the underfilled address is the end node in the prefix tree map model, and record the number of users whose underfilled addresses are end nodes. , The number of users whose addresses are less than the previous node of the end node ,like , the node optimization method is used to optimize the end node in the prefix tree map model.

2. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 1 is characterized by: The system connection module (100) applies to the cross-border e-commerce system and the map system for API information, request method, required request parameters and authentication method respectively; Then, the network request library in the programming language is used to construct the HTTP request, and the authentication information is added to the HTTP request. The constructed request is sent to the corresponding servers of the cross-border e-commerce system and the map system through the network. The request is transmitted in the form of data packets in the network in accordance with the format and rules specified by the HTTP protocol; After receiving the request, the server of the cross-border e-commerce system and the map system parses the request. If the request is legal, the server of the cross-border e-commerce system and the map system queries the pre-sale order data and map information according to the request content. The server of the cross-border e-commerce system and the map system respectively encapsulates the queried pre-sale order data and map information into a response message, and returns the response message in JSON format. The response message is also transmitted in the network in the form of a data packet in accordance with HTTP; After receiving the response message, the system connection module (100) uses a corresponding parsing library to convert the data into a data structure in a programming language according to the format of the response message, thereby acquiring the pre-sale order data and map information.

3. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 2 is characterized by: The prefix tree map construction module (210) constructs a prefix tree map model, specifically through the following steps: Step 1: Define the prefix tree node structure through the address hierarchical structure, each node contains character data, child node list and end mark information; Each character data of the prefix tree node structure represents a character in the address, the child node list stores all child nodes of the node, and the end mark indicates whether it is the end of a complete address; Step 2: Address splitting and insertion: Perceiving the address hierarchical structure predefined in the map system in the system connection module (100), traversing each address information in the map, splitting the addresses in the map information according to the address hierarchical structure, and sorting the address information according to the address hierarchical structure to determine the character data corresponding to each address; Start the operation from the root node of the prefix tree, and insert each character data in the address into the prefix tree in turn. During the insertion process, check the child node list of the current node. If there is a node corresponding to the character data, move the character data to the child node, and continue to insert the next character data according to the address hierarchy. If it does not exist, create a new node, add the character data to the child node list of the current node, and then move to the newly created node; Step 3: After traversing the numbers corresponding to the entire address, set the end mark of the end node to True, indicating that the path formed from the root node to this node represents a complete address.

4. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 3 is characterized by: The steps of the fuzzy matching algorithm in the address structure analysis module (220) are as follows: Step 1: respectively sense the hierarchical structure of the address filled in by the user and the address in the map information, combine the first-level addresses in the map information into a first address library, combine the second-level addresses into a second address library, until all the addresses in the map information are combined into an address library; Call out addresses corresponding to different hierarchical structures from the address filled in by the user, define them as the first address and the second address, until all the addresses in the filled-in address are defined; Step 2: Match the first address with the first-level address library. If the same address exists, the first-level address filled in by the user is determined to be correct. Otherwise, the similarity between the first-level address and all addresses in the first-level address library is calculated, and the address with the greatest similarity in the first-level address library is marked. Set a similarity threshold. If the maximum similarity is greater than the similarity threshold, the address filled in by the user will be replaced with the address with the highest similarity in the first-level address database. Step 3: After the first address is matched with the first-level address library, the second address is retrieved again to be matched with the second-level address library until all the addresses corresponding to the filled-in address are matched.

5. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 4 is characterized by: The method for calculating the similarity between the first-level address and all addresses in the first-level address library by the fuzzy matching algorithm in the address structure analysis module (220) is as follows: A method for establishing a connection with a cross-border e-commerce system through the system connection module (100), establishing a connection with a text corpus system, segmenting the text corpus, and assigning a row and column position in the matrix to each word; Calculate the co-occurrence of the first-level addresses in the text corpus and all the addresses in the first address library respectively. If the first-level address Y and an address O in the first address library appear in the same text, the value of the corresponding position in the co-occurrence matrix will increase by 1, and the vocabulary statistics of the first-level addresses and all the addresses in the first address library are constructed into a co-occurrence matrix that records the number of vocabulary co-occurrences. After obtaining the co-occurrence matrix, the first-level address and all addresses in the first address library correspond to a row vector or column vector in the matrix. The vector corresponding to the first-level address Y is , an address O in the first address library is , at this time, the similarity between vectors is calculated by the cosine similarity calculation method.

6. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 5 is characterized by: The address structure analysis module (220) inputs the filled-in address converted into the same format as the map information into the prefix tree map model constructed by the prefix tree map construction module (210), splits the filled-in address according to the address hierarchical structure, and determines the corresponding character data. , insert character data into the prefix tree map model; The perceptual prefix tree map model is ,in It is a list of different child nodes in the prefix tree map model, and the child node lists are different child nodes, , , , ; During insertion into the prefix tree map model: If a character data in the filled-in address does not have a corresponding node in the prefix tree map model, it is determined that there is missing address information in the user's filled-in address. At this time, the node without address information is skipped, and the sublist and next node list without corresponding nodes are called out in the prefix tree map model to match the unmatched character data in the filled-in address again; , it is determined that there is missing address information in the address filled in by the user ; The missing address information is filled in according to the address information adjacent to the missing address information in the address hierarchy structure.

7. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 6 is characterized by: In the address structure analysis module (220), if a character data of the filled-in address does not have a corresponding node in the prefix tree map model, it is determined that there is wrong address information in the address filled in by the user, and the wrong address information is deleted and converted into less-filled address information; like , it is determined that there is wrong address information in the address filled in by the user; If the end node is not True after all the character data of the filled-in address are inserted into the prefix tree map model, but the nodes and character data before it are exactly the same, it is also determined that there is missing address information in the address filled in by the user; , then determine the end node corresponding to the prefix tree map model abnormal; If the end node is True after all the character data of the filled-in address is inserted into the prefix tree map model, it is determined that the address filled in by the user is correct, such as , it is determined that the address filled in by the user is correct.

8. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 2 is characterized by: The address processing master control module (200) further comprises an address error correction and filling module (230), wherein the address error correction and filling module (230) is used to sense insufficiently filled addresses and incorrectly filled address information, fill insufficiently filled addresses and modify incorrectly filled address information according to the address hierarchical structure, and input the filled and modified address information into the cross-border e-commerce system through the connection established between the system connection module (100) and the cross-border e-commerce system, thereby prompting the user to confirm the address information.

9. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 1, characterized in that: The user interaction feedback module (300) senses the end node of the prefix tree map model Corresponding to the character data of the current user, and sensing the address information in the historical order data of the cross-border e-commerce system, all the historical address information is converted into corresponding character data, which is defined as a historical character data set; Match and calculate the current character data that is exactly the same as the historical character data set, and record the number of users ; Remove the end node from the historical character data set , match and calculate again and remove the end node Exactly the same current character data, record the number of users .

10. The cross-border e-commerce order management system based on pre-sale data analysis according to claim 9, characterized in that: The node optimization method in the user interaction feedback module (300) is , then the prefix tree map model end node in the address processing master control module (200) Remove the end node The previous node is used as the new node Otherwise, the end node optimization.

Images