Terminal logistics user identification method and device
By extracting feature and calculating similarity of express delivery bills, the problem of information fusion in terminal logistics is solved, accurate matching and efficient reuse are achieved, and user experience and cost control are improved.
Patent Information
- Application Number
- CN202510335658.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-20
- Publication Date
- 2025-07-04
AI Technical Summary
In the terminal distribution link of express logistics, the existing technology cannot effectively integrate multi-dimensional fragmented information, resulting in low matching accuracy, dynamic changes in user information, resulting in failure of notifications, declining user experience, and serious information island phenomenon.
Through the identification model, feature extraction of express delivery pages, calculate the similarity of pages feature vectors, and combine large models and third-party APIs to standardize and match information, so as to achieve accurate matching and efficient reuse of end logistics information.
It improves the accuracy of information matching, reduces costs, improves user experience, adapts to dynamic changes in user information, and reduces false alarms and missed reports.
Smart Images
Figure CN120260070A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of computer technology, and particularly to a method and device for identifying end - logistics users. Background Art
[0002] With the booming development of e - commerce, express logistics has become an indispensable part of modern life. In the end - delivery link of express logistics, accurately and efficiently matching user information is crucial for improving delivery efficiency, reducing costs, ensuring information security, and optimizing the user experience. However, in reality, the diversity of user information, the need for privacy protection, and the information - island phenomenon among logistics platforms have brought huge challenges to the effective reuse of information.
[0003] In the prior art, in the end - link of express logistics, relying on caching or a single information field for matching cannot solve the problem of multi - dimensional fragmented information fusion, and the matching accuracy is low; due to different input specifications, transportation damage, OCR recognition errors, etc. for express waybills from multiple platforms and channels, the same user may have multiple non - standard information, making the matching difficult; due to the frequent change of the user's mobile phone number and the continuous update of the address, etc., it is impossible to effectively cope with the dynamic changes of user information, resulting in problems such as notification failure and degraded user experience. Summary of the Invention
[0004] An object of this application is to provide a method and device for identifying end - logistics users, which can achieve accurate matching and efficient reuse of end - logistics information through the matching technology of waybill information on the express waybill, while taking into account cost control and user experience.
[0005] According to one aspect of this application, a method for identifying end - logistics users is provided. The method includes:
[0006] Determine an identification model for identifying the express waybill of the user;
[0007] Obtain the fragmented information on the current express waybill of the target user and process it into structured data;
[0008] Input the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain the waybill feature vector corresponding to the current express waybill of the target user;
[0009] Calculate the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user;
[0010] Determine the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user.
[0011] Further, in the above method, the identification model determined for identifying the user's express waybill includes:
[0012] Pre-training on a large-scale historical waybill dataset, where the pre-training steps include:
[0013] Data loading, for loading the processed user identification dataset;
[0014] Data annotation, for converting the user data in the user identification dataset into the dataset format of the BERT model, and splitting the large-scale historical waybill dataset into a training set, a test set, and a validation set;
[0015] Determine training parameters, where the training parameters include one or more of the learning rate, batch size, and number of training epochs;
[0016] Model training, for training the BERT model using the training set split from the historical waybill dataset.
[0017] Further, in the above method, the method further includes:
[0018] Obtain fragmented information of a new express waybill and perform data annotation;
[0019] Based on the fragmented information of the new express waybill after annotation, regularly adjust the training set, test set, and validation set respectively, and correct the training parameters;
[0020] Fine-tune and test the pre-trained identification model based on the corrected training parameters and the adjusted training set, test set, and validation set to obtain an updated identification model.
[0021] Further, in the above method, the obtaining fragmented information on the current express waybill of the target user and processing it into structured data includes:
[0022] Based on optical character recognition (OCR) by integrating a third-party SDK, collect the current express waybill of the target user and recognize a JSON-structured string information;
[0023] Extract the recognized JSON-structured string information to obtain the fragmented information of the current express waybill of the target user;
[0024] Process the fragmented information of the current express waybill of the target user to obtain the structured data of the current express waybill of the target user.
[0025] Further, in the above method, if the fragmented information includes the original user address, where processing the fragmented information of the current express waybill of the target user to obtain the structured data of the current express waybill of the target user includes:
[0026] Call a third-party map API, and send the fragmented information to the address parsing API of the third-party map, so that the address parsing API of the third-party map returns response data;
[0027] Extract the corrected address information from the returned response data;
[0028] Compare and update the corrected address information with the original user address in the fragmented information to obtain the corrected address information of the current express waybill of the target user, so as to obtain the structured data of the current express waybill of the target user.
[0029] Further, in the above method, the method further includes:
[0030] Determine the confidence level based on the similarity and the business logic method of the target user;
[0031] If the confidence level is greater than the first preset threshold, directly use the waybill information of the historical express waybill corresponding to the highest similarity for business processing;
[0032] If the confidence level is less than or equal to the first preset threshold and greater than the second preset threshold, initiate a user confirmation to the client of the target user, or initiate a collaborative confirmation to a third party associated with the target user;
[0033] If the confidence level is less than or equal to the second preset threshold, send a prompt message indicating untrustworthiness to the target user.
[0034] Further, in the above method, the method further includes:
[0035] Encrypt the information stored in the database;
[0036] When transmitting the information stored in the database, use TLS / SSL to encrypt the transmitted information;
[0037] Use role-based access control and fine-grained permission management to manage access to the information stored in the database.
[0038] Further, in the above method, the method further includes:
[0039] Write the structured data of the current express waybill of the target user into a structured database;
[0040] Store the waybill feature vector corresponding to the current express waybill of the target user based on BERT + Elasticsearch.
[0041] According to another aspect of the present application, there is also provided a non-volatile storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor, the processor is caused to implement the end logistics user identification method as described above.
[0042] According to another aspect of the present application, there is also provided a device for end logistics user identification, wherein the device includes:
[0043] One or more processors;
[0044] A computer-readable medium for storing one or more computer-readable instructions,
[0045] When the one or more computer-readable instructions are executed by the one or more processors, the one or more processors are caused to implement the end logistics user identification method as described above.
[0046] Compared with the prior art, the present application first determines an identification model for identifying a user's express waybill; in an actual application scenario, obtains fragmented information on the current express waybill of the target user and processes it into structured data; inputs the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain the waybill feature vector corresponding to the current express waybill of the target user; calculates the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user; and determines the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user. Through the waybill information matching technology on the express waybill, accurate matching and efficient reuse of end logistics information are achieved, while taking into account cost control and user experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, objects, and advantages of the present application will become more apparent:
[0048] Figure 1 A flowchart showing a method for end logistics user identification according to one aspect of the present application;
[0049] Identical or similar reference numerals in the drawings represent identical or similar components. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0050] The present application will be further described in detail below with reference to the drawings.
[0051] In a typical configuration of the present application, the terminal, the device of the service network, and the trusted party each include one or more processors (CPUs), an input / output interface, a network interface, and a memory.
[0052] The memory may include non-permanent memory in a computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.
[0053] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.
[0054] As Figure 1 shown, an aspect of the present application presents a schematic flow diagram of a method for identifying end-logistics users. This method is mainly based on user-related information on the express waybill, such as the user's name, user contact information, and the address related to the user's receipt and sending of packages, etc. This method includes steps S11, S12, S13, S14, and S15, where the specific implementation is as follows:
[0055] Step S11, determine an identification model for identifying the express waybill of the user;
[0056] Step S12, obtain the fragmented information on the current express waybill of the target user and process it into structured data;
[0057] Step S13, input the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain the waybill feature vector corresponding to the current express waybill of the target user;
[0058] Step S14, calculate the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user;
[0059] Step S15, determine the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user.
[0060] Through the above steps S11 to S15, first train and determine an identification model for identifying the express waybill of a user. Then, in an actual application scenario, after obtaining the fragmented information on the current express waybill of the target user and processing it into structured data, input the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain the waybill feature vector corresponding to the current express waybill of the target user; calculate the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user; determine the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user. Through the waybill information matching technology on the express waybill, accurate matching and efficient reuse of end - of - line logistics information are realized, while taking into account cost control and user experience.
[0061] In the embodiment of the present application, when performing efficient identification of end - of - line logistics users, it is mainly based on the fuzzy matching technology of a large model to process the fragmented information on the express waybill and improve the accuracy of information matching. Specifically, it includes the following steps: First, collect the fragmented information, such as the OCR information on the current express waybill, etc. Then, process the fragmented information of the current express waybill into structured data and use a pre - trained large model (such as BERT, etc.) to convert the structured data carrying the fragmented information into a waybill feature vector; Next, use a vector similarity calculation method (such as cosine similarity, etc.) to perform fuzzy matching on the waybill feature vector of the current express waybill of the target user, calculate the similarity from the waybill feature vectors of the historical express waybills of the target user, and find the most similar record, that is, the waybill information of the historical express waybill corresponding to the highest similarity; Finally, determine the waybill information of the historical express waybill corresponding to the highest similarity obtained by fuzzy matching as the standardized waybill information of the current express waybill of the target user, so as to generate complete and consistent waybill data for subsequent business scenarios such as message push, service matching, and data analysis and statistics after user information identification, realizing accurate matching and efficient reuse of end - of - line logistics information, while taking into account cost control and user experience. Among them, in the present application, through multi - dimensional feature fusion and large - model reasoning, the user identity can be accurately identified, thereby reducing false alarms and missed reports.
[0062] Continuing with the above embodiments of the present application, when training an identification model for identifying the express waybill of a user, based on a large model of the Transformer architecture, pre-trained on a large-scale text dataset, feature vector matching and scoring are performed. Among them, when determining the identification model for identifying the express waybill of a user in step S11, it specifically includes:
[0063] Perform pre-training on a large-scale historical waybill dataset. Among them, the pre-training steps include:
[0064] Data loading, used to load and process the user identification dataset;
[0065] Data annotation, used to convert the user data in the user identification dataset into the dataset format of the BERT model, and divide the large-scale historical waybill dataset into a training set, a test set, and a validation set;
[0066] Determine training parameters, where the training parameters include one or more of the learning rate, batch size, and number of training epochs;
[0067] Model training, used to train the BERT model using the training set segmented from the historical waybill dataset, so as to obtain an identification model for identifying the express waybill of a user, thereby realizing the training of the large model.
[0068] When creating an Elasticsearch index, define a mapping containing a dense vector field, store the waybill information feature vectors of users in the historical waybill dataset into the Elasticsearch index, use Elasticsearch for similarity search, find the most similar waybill feature vectors by calculating the cosine similarity. Finally, the scoring mechanism uses similarity calculation methods such as chord similarity to calculate the similarity between the waybill feature vectors of the current express waybill and the waybill feature vectors of the historical express waybills, and scores according to the similarity. The higher the similarity, the higher the score, thereby realizing the effective search and matching of the waybill information of the current express waybill.
[0069] Continuing with the above embodiments of the present application, in an actual application scenario, with continuous data updates, such as changes in the user's mobile phone number, address updates, etc., the model can be fine-tuned to adapt to the dynamic changes of user information. Among them, the specific implementation steps for fine-tuning the model are:
[0070] First, obtain the fragmented information of the new express waybill and perform data annotation. Here, the data annotation of the fragmented information of the obtained new express waybill can be real-time, periodic, or regular. The new express waybill can be a new form of express waybill with a specific address or waybill format, etc.;
[0071] Then, based on the fragmented information of the new express waybill after annotation, the training set, test set, and validation set for training and validating the recognition model are adjusted regularly, and the training parameters are corrected according to the effects of training, testing, and validation;
[0072] Finally, based on the corrected training parameters and the adjusted training set, test set, and validation set, the pre-trained recognition model is fine-tuned and tested, so as to obtain an updated recognition model, achieving the purpose of updating the recognition model, thus being more adaptable to the constantly changing express waybill and having a higher user experience.
[0073] Continuing with the above embodiments of the present application, when the step S12 obtains the fragmented information on the current express waybill of the target user and processes it into structured data, it specifically includes:
[0074] By integrating a third-party SDK, based on optical character recognition (OCR), the current express waybill of the target user is collected, and a JSON-structured string information is recognized; here, the information collection of OCR is carried out by integrating a third-party SDK to obtain a JSON-structured string information;
[0075] The JSON-structured string information recognized is extracted to obtain the fragmented information of the current express waybill of the target user; for example, the user name, user contact information, barcode, and the address of the user's sending and receiving are parsed;
[0076] The fragmented information of the current express waybill of the target user is processed; for example, the parsing of the address in the current express waybill needs to be processed into structured data (such as detailed to the house number, etc. like community, school, village, street, etc.), to obtain the structured data of the current express waybill of the target user.
[0077] It should be noted that the structured address in the structured data refers to the address in standard format / province / city / district / community, village, school, business district / building / unit / house number. The latter part of the address for sending and receiving of the express waybill is usually filled in freely, and combined with the error of OCR recognition, so it needs to be corrected.
[0078] Pre-train a large model on a large-scale historical address dataset to learn various writing methods of different addresses and their standardized forms, and combine various writing methods, historical address data, and a third-party map API to standardize and correct the structured address information. Further, if the fragmented information includes the original user address, where the processing of the fragmented information of the current express waybill of the target user to obtain the structured data of the current express waybill of the target user specifically includes:
[0079] Extract unstandardized address information from the historical address dataset or call a third-party map API for the unstandardized address information corresponding to the current express waybill, and send the fragmented information to the address parsing API of the third-party map so that the address parsing API of the third-party map returns response data;
[0080] Extract the corrected address information from the returned response data.
[0081] Compare and update the corrected address information with the original user address in the fragmented information to obtain the corrected address information of the current express waybill of the target user, so as to obtain the structured data of the current express waybill of the target user. Among them, the structured data of the current express waybill of the target user not only includes structured address information, but also includes user-related information such as structured user name and structured user contact information, which not only realizes address standardization, but also realizes the correction of structured information on the user's express waybill.
[0082] In this scenario, a BERT model based on Transformer is used. The dataset includes daily corrected address information and data with various writing rules. The daily dataset is supplemented by triggering training regularly through the CronJon scheduling system, and the trained model is updated online for actual business applications.
[0083] Following the above embodiments of the present application, after determining the standardized waybill information of the current express waybill of the target user in step S15, it can be customized and used according to different business scenarios. Among them, the method further includes:
[0084] Determine the confidence level based on the similarity and the business logic method of the target user; here, the confidence level is determined in the following way: after feature vectorization, calculate according to the similarity between feature vectors. Here, cosine similarity is used for calculation. In addition, some metrics will be calculated in combination with the business logic method of the target user for assistance, so as to jointly obtain the confidence level;
[0085] In the process of applying the confidence level, if the confidence level is greater than the first preset threshold, directly use the waybill information of the historical express waybill corresponding to the highest similarity for business processing;
[0086] If the confidence level is less than or equal to the first preset threshold and greater than the second preset threshold, initiate user confirmation to the client of the target user, or initiate collaborative confirmation to a third party associated with the target user;
[0087] If the confidence level is less than or equal to the second preset threshold, send a prompt message indicating untrustworthiness to the target user.
[0088] In this embodiment, the notification method is dynamically selected according to the matching confidence, reducing the notification cost of the service logic layer.
[0089] Continuing with the above embodiments of the present application, strict compliance measures and security audit mechanisms also need to be taken for the information stored in the database to ensure that the system complies with relevant laws and regulations and reduces compliance risks. For example:
[0090] Encrypt the information stored in the database; for example, encrypt sensitive information such as user addresses, user contact information, and user sending and receiving addresses stored in the database to ensure the security during data storage and prevent data leakage.
[0091] When transmitting the information stored in the database, such as API calls, data backups, etc., use TLS / SSL to encrypt the transmitted information to ensure security during data transmission and prevent data leakage.
[0092] Use role-based access control and fine-grained permission management to manage access to the information stored in the database to ensure that only authorized users can access sensitive information in the database, etc.
[0093] It is also possible to record all user operations and system events to ensure traceability.
[0094] Continuing with the above embodiments of the present application, the method further includes:
[0095] Write the structured data of the current express waybill of the target user into a structured database to ensure the accuracy and integrity of the data, where the structured database includes but is not limited to a MySQL database, etc.;
[0096] Store the waybill feature vector corresponding to the current express waybill of the target user based on bert+elasticsearch, so as to realize the storage of the waybill feature vector corresponding to the current express waybill of the target user, thereby enriching the database of the waybill feature vectors of historical express waybills.
[0097] According to another aspect of the present application, a non-volatile storage medium is further provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor, the processor is enabled to implement the end logistics user identification method as described above.
[0098] According to another aspect of the present application, a device for end logistics user identification is further provided, where the device includes:
[0099] One or more processors;
[0100] A computer-readable medium for storing one or more computer-readable instructions,
[0101] when the one or more computer-readable instructions are executed by the one or more processors, the one or more processors implement the end logistics user identification method as described above.
[0102] Herein, for the detailed content of each embodiment in the device for end logistics user identification, reference may be specifically made to the corresponding part of the embodiment of the end logistics user identification method described above, and details are not repeated herein.
[0103] In summary, the present application first determines an identification model for identifying the express waybill of a user; in an actual application scenario, obtains fragmented information on the current express waybill of a target user and processes it into structured data; inputs the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain a waybill feature vector corresponding to the current express waybill of the target user; calculates the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user; determines the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user, and realizes the accurate matching and efficient reuse of end logistics information through the waybill information matching technology on the express waybill, while taking into account cost control and user experience.
[0104] It should be noted that the present application can be implemented in software and / or a combination of software and hardware. For example, it can be implemented using an application specific integrated circuit (ASIC), a general purpose computer, or any other similar hardware device. In one embodiment, the software program of the present application can be executed by a processor to implement the steps or functions described above. Similarly, the software program (including related data structures) of the present application can be stored in a computer-readable recording medium, such as a RAM memory, a magnetic or optical drive, or a floppy disk and the like. In addition, some steps or functions of the present application can be implemented using hardware, for example, as a circuit that cooperates with a processor to execute each step or function.
[0105] In addition, a part of this application can be applied as a computer program product, such as computer program instructions. When executed by a computer, through the operation of this computer, it can call or provide the methods and / or technical solutions according to this application. The program instructions for calling the methods of this application may be stored in a fixed or removable recording medium, and / or transmitted through a data stream in a broadcast or other signal-bearing medium, and / or stored in the working memory of a computer device running according to the program instructions. Here, an embodiment according to this application includes a device, and this device includes a memory for storing computer program instructions and a processor for executing the program instructions. Among them, when the computer program instructions are executed by the processor, it triggers the device to run based on the methods and / or technical solutions according to the foregoing multiple embodiments of this application.
[0106] For those skilled in the art, it is obvious that this application is not limited to the details of the above-mentioned exemplary embodiments, and without departing from the spirit or basic characteristics of this application, this application can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of this application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in this application. Any reference signs in the claims should not be regarded as limiting the claimed rights. In addition, obviously, the word "including" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or devices stated in the device claims can also be implemented by one unit or device through software or hardware. First, second, etc. are used to represent names and do not represent any specific order.
Claims
1. A method for identifying end - logistics users, wherein, The method includes: Determining an identification model for identifying the user's express waybill; Obtaining the fragmented information on the current express waybill of the target user and processing it into structured data; Inputting the structured data of the current express waybill of the target user into the identification model for feature extraction to obtain the waybill feature vector corresponding to the current express waybill of the target user; Calculating the similarity between the waybill feature vector corresponding to the current express waybill of the target user and the waybill feature vectors of at least one historical express waybill of the target user; Determining the waybill information of the historical express waybill corresponding to the highest similarity as the standardized waybill information of the current express waybill of the target user.
2. The method according to claim 1, wherein The determining of the identification model for identifying the user's express waybill includes: Performing pre-training on a large-scale historical waybill dataset, where the pre-training steps include: Data loading, for loading the processed user identification dataset; Data annotation, for converting the user data in the user identification dataset into the dataset format of the BERT model and splitting the large-scale historical waybill dataset into a training set, a test set, and a validation set; Determining training parameters, where the training parameters include one or more of a learning rate, a batch size, and the number of training epochs; Model training, for training the BERT model using the training set split from the historical waybill dataset.
3. The method according to claim 2, wherein, The method further includes: Obtaining the fragmented information of a new express waybill and performing data annotation; Regularly adjusting the training set, the test set, and the validation set respectively based on the fragmented information of the new express waybill after annotation, and correcting the training parameters; Fine-tuning and testing the pre-trained identification model based on the corrected training parameters and the adjusted training set, test set, and validation set to obtain an updated identification model.
4. The method according to claim 1, wherein The obtaining of the fragmented information on the current express waybill of the target user and processing it into structured data includes: Collecting the current express waybill of the target user based on optical character recognition (OCR) by integrating a third-party SDK and identifying a JSON-structured string information; Extracting the identified JSON-structured string information to obtain the fragmented information of the current express waybill of the target user; Processing the fragmented information of the current express waybill of the target user to obtain the structured data of the current express waybill of the target user.
5. The method according to claim 3 or 4, wherein If the fragmented information includes the original user address, where the processing of the fragmented information of the current express waybill of the target user to obtain the structured data of the current express waybill of the target user includes: Invoking a third-party map API and sending the fragmented information to the address resolution API of the third-party map so that the address resolution API of the third-party map returns response data; Extracting the corrected address information from the returned response data; Compare and update the corrected address information with the original user address in the fragmented information to obtain the corrected address information of the current express waybill of the target user, so as to obtain the structured data of the current express waybill of the target user.
6. The method according to claim 1, wherein The method further includes: Determine the confidence level based on the similarity and the business logic method of the target user; If the confidence level is greater than the first preset threshold, directly use the waybill information of the historical express waybill corresponding to the highest similarity for business processing; If the confidence level is less than or equal to the first preset threshold and greater than the second preset threshold, initiate user confirmation to the client of the target user, or initiate collaborative confirmation to a third party associated with the target user; If the confidence level is less than or equal to the second preset threshold, send a prompt message indicating untrustworthiness to the target user.
7. The method according to claim 5, wherein The method further includes: Encrypt the information stored in the database; When transmitting the information stored in the database, use TLS / SSL to encrypt the transmitted information; Use role-based access control and fine-grained permission management to manage access to the information stored in the database.
8. The method according to claim 5, wherein The method further includes: Write the structured data of the current express waybill of the target user into the structured database; Store the waybill feature vector corresponding to the current express waybill of the target user based on bert+elasticsearch.
9. A non-volatile storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor, the processor implements the method according to any one of claims 1 to 8.
10. An apparatus for identifying end - logistics users, wherein, The device includes: One or more processors; A computer-readable medium for storing one or more computer-readable instructions, When the one or more computer-readable instructions are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 8.