Subway operation data integration sharing method and system and electronic equipment

Abstract

The invention provides a subway operation data integration and sharing method and system and electronic equipment, and relates to the technical field of data processing, and the method comprises the steps: obtaining distributed subway data ports, and generating a communication topology in combination with an operation data interaction transmission network; data types of subway operation are interacted, data dichotomy is carried out, and data classification identifiers are determined; constructing a data processing model according to the data classification identifier, decoupling the data processing model, and arranging the data processing model at a subway data port based on communication topology; and returning the subway operation data, generating confusion data based on a first generator of a data transmission end to perform insertion confusion desensitization on the sensitive data, transmitting the sensitive data to a data receiving end, and performing confusion discrimination and data restoration based on a second discriminator. The technical problem that in the prior art, privacy disclosure risks exist in metro operation data management is solved, and the technical effect of improving data security in the metro operation data sharing process is achieved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method, system and electronic equipment for integrating and sharing subway operation data. Background Art

[0002] In modern subway operation systems, with the widespread application of various intelligent devices and information technologies, the management of subway operation data has become a key part of operation management. Existing subway operation data management methods mainly focus on centralized data collection and basic storage management, and transmit the data generated by various subsystems, such as ticketing systems, train operation monitoring systems, equipment maintenance systems, etc., to a centralized data management center. However, this method has significant deficiencies in data security.

[0003] First, there is a lack of effective classification mechanism in terms of data privacy protection. For example, passengers' personal information (such as identity information, travel trajectory, etc.) and train real-time operation status data are collected and stored equally, which leads to a greater risk of leakage of these privacy and sensitive information in the data sharing link. Secondly, there is a lack of refined means to control sensitive information, and there is no corresponding protection mechanism for sensitive data. Once these data are leaked, it will not only damage the personal rights and interests of passengers, but also endanger the safety and stability of subway operations.

[0004] In summary, the protection mechanism for sensitive data in the existing subway operation data management methods is not perfect and cannot effectively deal with the risk of data privacy leakage. Summary of the invention

[0005] The present application provides a method, system and electronic device for integrating and sharing subway operation data, which solves the technical problems in the prior art of subway operation data management that lack fine control over data privacy and sensitive information and have the risk of privacy leakage, and achieves the technical effect of improving data security in the process of sharing subway operation data.

[0006] In view of the above problems, on the one hand, the present application provides a method for integrating and sharing subway operation data, the method comprising: obtaining distributed subway data ports, combining the transmission network for interactive operation data, generating a communication topology, the subway data port being a multi-line data terminal connected to a subway operation information platform; interactive subway operation data types, performing data binary classification, and determining data classification identifiers, wherein the binary classification standard is a preset sensitivity; constructing a data processing model with the data classification identifier, decoupling the data processing model, and deploying it on the subway data port based on the communication topology; wherein constructing the data processing model comprises: identifying the data classification identifier, determining true sample data based on sensitive data types, and training a first generator and a second discriminator based on alternating iterations of an adversarial network; wherein the first generator and the second discriminator are trained on the data processing model based on the data classification identifier; The generator is initialized with parameter assignment, and a random noise vector is introduced as the input of the first generator to determine the generated data once; in combination with the second discriminator, the generated data once is compared with the true sample data to determine the first judgment data; based on the first judgment data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met to obtain a data processing model; wherein the adversarial equilibrium condition is that the generated data of the first generator meets the preset authenticity, and the second discriminator meets the preset judgment probability for the generated data of the first generator; the first generator and the second discriminator are split as a decoupling method; the subway operation data is transmitted back, and the obfuscated data generated by the first generator at the data transmission end is inserted into the sensitive data for obfuscation and desensitization, and transmitted to the data receiving end, and the obfuscation judgment and data restoration are performed based on the second discriminator.

[0007] On the other hand, the present application also provides a subway operation data integration and sharing system, which includes: a communication topology generation module, which is used to obtain distributed subway data ports, and generate a communication topology in combination with a transmission network for interactive operation data, wherein the subway data port is a multi-line data terminal connected to a subway operation information platform; a data classification module, which is used to interact with the data type of subway operation, perform data binary classification, and determine a data classification identifier, wherein the binary classification standard is a preset sensitivity; a data processing model construction module, which is used to construct a data processing model with the data classification identifier, decouple the data processing model, and deploy it on the subway data port based on the communication topology; wherein constructing the data processing model includes: identifying the data classification identifier, determining true sample data based on the sensitive data type, and training the first generator and the second discriminator based on alternating iterations of the adversarial network. ; wherein, the first generator and the second discriminator are initialized with parameter assignment, a random noise vector is introduced as the input of the first generator, and the generated data is determined once; in combination with the second discriminator, the generated data and the true sample data are judged to determine the once determined data; based on the once determined data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met to obtain a data processing model; wherein the adversarial equilibrium condition is that the generated data of the first generator meets the preset authenticity, and the second discriminator meets the preset discrimination probability for the generated data of the first generator; the splitting of the first generator and the second discriminator is used as a decoupling method; a data interactive sharing management module, the data interactive sharing management module is used to transmit subway operation data back, and the obfuscated data generated by the first generator at the data transmission end is inserted into the sensitive data for obfuscation and desensitization, and transmitted to the data receiving end, and the obfuscation discrimination and data restoration are performed based on the second discriminator.

[0008] In a third aspect, the present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in the subway operation data integration and sharing method described in any one of the above claims are implemented.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages: Obtain distributed subway data ports, combine with the transmission network for interactive operation data, generate communication topology, and build a basic network architecture for subsequent data processing, interactive sharing and other operations. The data type of interactive subway operation is classified according to the preset sensitivity and the data classification mark is determined, so that the subsequent data processing can be targeted according to the different properties (sensitivities) of the data, laying the foundation for building a suitable data processing model. Based on the data classification mark, a data processing model is constructed based on an adversarial network, including a first generator and a second discriminator. This model structure can effectively process subway operation data. The data processing model is decoupled and arranged on the subway data port based on the communication topology, so as to utilize the previously established communication topology structure to perform effective data processing on each data port. The subway operation data is transmitted back. During the data transmission process, the data is desensitized based on the first generator at the data transmission end to protect data security; the data is restored based on the second discriminator at the data receiving end to ensure data availability. Combined with the communication topology, data interactive sharing management is performed to achieve secure sharing of subway operation data.

[0010] In summary, this application builds a complete subway operation data integration and sharing system, from the construction of basic communication topology, to the detailed classification of data, to the construction and layout of the core data processing model, and finally to the interactive sharing and management of data. This solution effectively improves the security of subway operation data, protects sensitive data through data desensitization and restoration mechanisms, improves the efficiency and accuracy of data management, and allows different types of data to be properly processed; at the same time, it also enhances the effectiveness of data interaction and sharing, and provides accurate and timely data support for all aspects of subway operations.

[0011] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flowchart of the subway operation data integration and sharing method provided in an embodiment of the present application.

[0013] Figure 2 A schematic diagram of the structure of the subway operation data integration and sharing system provided in an embodiment of the present application.

[0014] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0015] Explanation of the reference numerals: communication topology generation module 10 , data classification module 20 , data processing model construction module 30 , data interaction and sharing management module 40 , bus 300 , receiver 301 , processor 302 , transmitter 303 , memory 304 , bus interface 305 . DETAILED DESCRIPTION

[0016] The embodiments of the present application provide a method, system and electronic device for integrating and sharing subway operation data. By building a communication topology and classifying data according to data sensitivity, and then constructing a data processing model based on an adversarial network, the interactive sharing management of subway operation data is completed, which solves the technical problems in the prior art that subway operation data management lacks fine control over data privacy and sensitive information and has the risk of privacy leakage, and achieves the technical effect of improving data security in the process of sharing subway operation data.

[0017] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a method for integrating and sharing subway operation data, the method comprising: Step S1: Obtain distributed subway data ports, combine with the transmission network for operation data interaction, and generate a communication topology. The subway data port is a multi-line data terminal connected to the subway operation information platform.

[0018] Specifically, the subway data port is a multi-line data port connected to the subway operation information platform. These ports are responsible for collecting data from different lines and different business systems. The subway operation information platform is an integrated system used to manage various types of subway operation information, including scheduling, maintenance, passenger information, etc. The communication topology is a network structure diagram composed of data ports and data transmission networks, which describes the connection relationship between each subway data port and the transmission network for operation data interaction and the path of data flow.

[0019] Transmission network for interactive operation data: This is a network system used to transmit subway operation data. It can be a wired or wireless network composed of various network devices (such as routers, switches, etc.). For example, in a subway system of a large city, there may be an optical fiber network to transmit train operation monitoring data, a wireless sensor network to transmit equipment status data in the carriage, etc. This network is responsible for transmitting and interacting data from various data ports.

[0020] First, determine all multi-line data terminals connected to the subway operation information platform, that is, subway data ports. These ports are scattered in various links of subway operation, such as ticket management, train scheduling, equipment monitoring, etc. Then, obtain the network architecture for transmitting and exchanging subway operation data between different data ports, that is, the transmission network. The transmission network can be a wired or wireless network composed of various network devices (such as routers, switches, etc.), which is responsible for transmitting and interacting data from each data port. For example, in a subway system in a large city, train operation monitoring data is transmitted through an optical fiber network, and equipment status data in the carriage is transmitted through a wireless sensor network. Next, combine these subway data ports with the transmission network for operation data interaction, analyze the connection relationship of each data port in the network, determine the relationship between nodes (data ports) and links (network connection lines), and use network topology tools to simulate network connections to generate communication topology.

[0021] By generating the communication topology, a basic network architecture is built for subsequent data processing, interactive sharing and other operations, and the basic framework of data transmission is determined.

[0022] Step S2: Interact with the data type of subway operation, perform data binary classification, and determine the data classification identifier, wherein the binary classification standard is a preset sensitivity.

[0023] Specifically, the data types of subway operations refer to various types of data generated during the operation of the subway. For example, ticketing data (such as ticket prices, number of tickets sold, etc.), train operation data (such as train speed, arrival time, etc.), equipment maintenance data (such as elevator maintenance records, track inspection data, etc.) and passenger flow data (the number of people entering and leaving the station at different time periods, etc.). Data classification identification is an identification used to mark the data classification results. It can be a simple number, letter or specific symbol, etc., used to distinguish between two types of data with different sensitivities. For example, "1" can be used to represent a type of data with high sensitivity, and "0" can be used to represent a type of data with low sensitivity; or "A" can be used to represent one type and "B" can be used to represent another type. The preset sensitivity is a pre-set standard for measuring the sensitivity of data, which is used to determine whether the data is sensitive data.

[0024] Extract all relevant subway operation data from various systems of subway operation (ticketing system, train control system, etc.), identify and classify different data types. Then, according to the preset sensitivity standards, divide the data into two categories: sensitive data and non-sensitive data. This process can use data analysis tools, such as the pandas library in Python, to classify the data by setting conditional judgment statements. After the classification is completed, a classification identifier is assigned to each type of data according to its sensitivity so that it can be identified and managed in subsequent processing. For example, the preset sensitivity stipulates that data involving passenger privacy (such as name, contact information) is highly sensitive data, and all data containing this type of information is classified into one category and marked as "sensitive data"; while data with less impact on privacy and security, such as the distribution of public facilities at subway stations, is classified into another category and marked as "non-sensitive data."

[0025] By classifying operational data into two categories and determining data classification identifiers, different types of data can be automatically identified during subsequent data processing, and different security measures can be taken for data of different sensitivities to improve the accuracy of data processing.

[0026] Step S3: Construct a data processing model with the data classification identifier, decouple the data processing model, and deploy it at the subway data port based on the communication topology, wherein constructing the data processing model includes: identifying the data classification identifier, determining the true sample data based on the sensitive data type, and training the first generator and the second discriminator based on the alternating iteration of the adversarial network; wherein, the first generator and the second discriminator are initialized with parameter assignments, and a random noise vector is introduced as the input of the first generator to determine the generated data once; in combination with the second discriminator, the generated data and the true sample data are determined to determine the first judgment data; based on the first judgment data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met to obtain the data processing model; wherein the adversarial equilibrium condition is that the generated data of the first generator meets the preset authenticity, and the second discriminator meets the preset judgment probability for the generated data of the first generator; the splitting of the first generator and the second discriminator is used as the decoupling method.

[0027] Specifically, according to the data classification identification determined in step S2, a data processing model is constructed based on an adversarial network to process and protect sensitive data. The data processing model consists of two parts: a first generator and a second discriminator. Among them, the first generator is used to generate desensitized data to ensure that sensitive information is protected; the second discriminator is used to determine whether the generated data is valid to ensure its accuracy during data restoration. Exemplarily, a deep learning framework (such as TensorFlow or PyTorch) can be used to build and train a data processing model. For the first generator, a multi-layer perceptron structure in a neural network can be used to generate some simulated data based on the characteristics of sensitive data; for the second discriminator, a convolutional neural network structure can be used to accurately determine the authenticity of the data.

[0028] Then, the constructed data processing model is decoupled so that each part can be operated, analyzed or optimized independently. The first generator is responsible for generating operation simulation data at the data transmission end, and the second discriminator is responsible for judging whether the received data is real data at the data receiving end. Based on the communication topology generated in step S1, the data processing model is deployed on the subway data port using network deployment tools (such as Docker containers, etc.) to ensure that the model can effectively process subway operation data.

[0029] To build a data processing model, we first read the data classification identifier of the operational data to identify the type of sensitive data, such as the passenger's name, ID number, etc. According to the type of sensitive data, we select real and unprocessed sensitive data from the original operational data as true sample data. These true sample data will be used to train the data processing model.

[0030] Using a deep learning framework (such as TensorFlow or PyTorch), with real sample data as the training data set, alternately iterate the training of the first generator and the second discriminator. At the beginning of the training, the first generator randomly generates some data (for example, for passenger privacy data, some virtual data with similar structure is generated), and then inputs these generated data into the second discriminator for discrimination. The second discriminator judges the authenticity of these generated data based on the characteristics of the real sample data and outputs the discrimination result. According to this discrimination result, adjust the parameters of the first generator. For example, when the second discriminator finds that the generated data is very different from the real sample data in a certain feature, it adjusts the weight of the neural network in the first generator through the back propagation algorithm. Then the next round of generation and discrimination is carried out, and the training is repeated alternately iteratively.

[0031] The first generated data is the data generated for the first time by the first generator based on the initialization parameters and the random noise vector input. The first determined data is the data result obtained by the second discriminator after determining the first generated data and the true sample data.

[0032] In the alternating iterative training process, first, the parameters of the first generator and the second discriminator are initialized. For example, when building these two models, the weights and biases of the neural network layers in the model can be assigned initial values ​​by defining the model structure and using the default initialization method (such as Xavier initialization). Then, a random noise vector is introduced as the input of the first generator to introduce a certain degree of randomness, diversify the generated data, and avoid limiting the trend from the beginning. This random noise vector can be generated by a random number generation library in Python (such as numpy). After receiving the random noise vector, the first generator generates data once according to its own neural network structure and initialization parameters.

[0033] The first generated data and the true sample data are input into the second discriminator for determination to obtain first determined data, which is a probability value indicating the possibility that the first generated data is true data.

[0034] Define a loss function (such as the cross entropy loss function), calculate the value of the loss function based on the difference between the first judgment data and the real sample, and then use the back propagation algorithm to adjust the parameters of the second discriminator based on this loss value. At the same time, according to the judgment result of the second discriminator, adjust the parameters of the first generator, and train the first generator again. This process will be repeated, and data will be regenerated, judged, and parameters adjusted each time until the adversarial equilibrium condition is met.

[0035] The adversarial equilibrium condition is a condition for judging whether the training of the first generator and the second discriminator has reached an ideal state. The preset authenticity is a pre-set standard for measuring the authenticity of the data generated by the first generator. The preset discrimination probability is a pre-set accuracy standard for the second discriminator to discriminate the data generated by the first generator. After each round of training is completed, check whether the first generator and the second discriminator meet the adversarial equilibrium condition, that is, whether the data generated by the first generator meets the preset authenticity standard, and whether the data generated by the second discriminator meets the preset discrimination probability. Exemplarily, it can be judged whether the preset authenticity is met by calculating the similarity between the data generated by the first generator and the true sample data in each feature; and whether the preset discrimination probability is met by counting the proportion of correct judgments made by the second discriminator in multiple discriminations. When the generated data of the first generator meets the preset authenticity, and the second discriminator meets the preset discrimination probability for the generated data of the first generator, it is considered that the training has reached the adversarial equilibrium condition, and the model obtained at this time is the final data processing model. This data processing model can be used for the subsequent processing of subway operation data to ensure data security when sharing data.

[0036] Step S4: Transmit the subway operation data back, generate obfuscated data based on the first generator at the data transmission end, insert obfuscation and desensitize the sensitive data, transmit it to the data receiving end, and perform obfuscation discrimination and data restoration based on the second discriminator.

[0037] Specifically, data desensitization refers to processing sensitive data to remove or hide some sensitive information without losing data features. For example, replacing the middle digits of a passenger's ID number with asterisks. Data restoration corresponds to data desensitization, which means restoring the desensitized data to a state that can be used normally at the data receiving end.

[0038] First, the subway operation data is transmitted back from each data port to the subway operation information platform. In this process, the data will be interactively shared and managed through the communication topology established in step S1. For example, the communication topology is a star communication topology, with the subway operation information platform as the central node, and the data of each subsystem will flow to the platform, and then shared with other related systems as needed. For example, the data of the ticketing system may be transmitted to the subway operation information platform first, and then the platform will share part of the data according to the needs of various departments. During the data sharing process, the first generator at the data transmission end desensitizes the data to be transmitted. For example, a regular expression is used to replace the passenger's name with a specific symbol or anonymous identifier. The second discriminator at the data receiving end identifies the desensitized data transmitted from the data transmission end, and restores the data to a state that can be used normally according to the restoration logic in the model.

[0039] The data desensitization and restoration mechanism ensures the security of sensitive data transmission and availability after receiving, thereby improving the efficiency and security of data management.

[0040] Furthermore, step S3 of the embodiment of the present application also includes: S31: Construct a discriminant loss function and a generative loss function, wherein the discriminant loss function is the cross entropy loss of the upper generated sample and the true sample data, and the generative loss function is the negative log-likelihood of the true probability of the upper discriminant result.

[0041] S32: Based on the discriminant loss function, adjust the parameter assignment of the second discriminator after one training, and determine the secondary parameter assignment, wherein minimizing the loss is the adjustment direction.

[0042] S33: Temporarily fix the secondary parameter assignment, and perform parameter assignment adjustment and secondary generation on the first generator.

[0043] Specifically, the loss function includes the discriminant loss function and the generation loss function. The discriminant loss function is a function used to measure the difference between the discrimination result of the second discriminator and the actual situation (the upper generated sample and the true sample data). The cross entropy loss is used here. The cross entropy loss is used to measure the difference between the predicted probability distribution and the true probability distribution. In the discrimination task, it can well reflect the degree of deviation between the discriminant effect of the discriminator and the actual situation. The generation loss function is a function used to measure the effect of the data generated by the first generator. It is defined here as the negative log-likelihood of the true probability of the upper discriminant result. The negative log-likelihood is used to measure the probability that the data generated by the generator is correctly identified as true data by the discriminator. This function can be used to evaluate the quality of the data generated by the generator.

[0044] After one training, the discriminant loss function value is calculated, and the parameters of the second discriminator are updated through the back propagation algorithm so that the discriminant loss function develops in the direction of minimization, thereby obtaining the secondary parameter assignment, that is, the parameter state of the second discriminator after adjustment.

[0045] Temporarily fix the secondary parameter assignment, that is, when adjusting the parameters of the first generator, temporarily keep the parameters of the second discriminator unchanged. According to the generated loss function, the optimizer of the first generator is used to adjust the parameters of the first generator through a similar back-propagation algorithm. After adjusting the parameters, the first generator generates data again (secondary generation), and this generated data will enter the next round of training and judgment cycle until the adversarial equilibrium condition is met.

[0046] Through alternating iterative training, the parameters of the first generator and the second discriminator are continuously adjusted to improve the performance of the data processing model and achieve effective data desensitization and secure sharing.

[0047] Furthermore, step S3 of the embodiment of the present application also includes: S34: Decouple the data processing model, and split the first generator and the second discriminator.

[0048] S35: traversing the communication topology, and determining the data port type of each subway data port based on the data interaction characteristics, wherein the data port type includes an output port, a receiving port, and a bidirectional port.

[0049] S36: Based on the data port type, perform port configuration on the first generator and the second discriminator.

[0050] Specifically, the data processing model is decoupled, the first generator and the second discriminator are split, and the generator and the discriminator are run as independent modules. The communication topology is traversed to access each data port. For each data port, its data interaction characteristics are analyzed, that is, the direction of data transmission of each port is unidirectional output, unidirectional reception or bidirectional transmission, so as to determine whether each data port belongs to the output end, the receiving end or the bidirectional end. The output end is only responsible for outputting data; the receiving end is responsible for receiving data; the bidirectional end can perform bidirectional data interaction, which can both output data and receive data.

[0051] The first generator and the second discriminator are set according to the data port type so that they can adapt to the data processing requirements of different types of ports. The first generator is configured at the output end to perform operations such as desensitization processing on the data to be output. The second discriminator is configured at the receiving end to perform operations such as restoration and verification on the received data. For the bidirectional end, the first generator and the second discriminator are configured, the first generator processes the data sent outward, and the second discriminator processes the received data.

[0052] According to the characteristics of data interaction, a corresponding data processing model is configured for each port to facilitate the smooth operation of the entire data processing and sharing process and ensure the effective management and protection of data.

[0053] Furthermore, step S4 of the embodiment of the present application also includes: S41: Determine a first transmission end and a second reception end for the returned subway operation data, wherein the first transmission end and the second reception end include at least one subway data port.

[0054] S42: Based on the data classification identifier, scan the returned subway operation data to identify sensitive data parts.

[0055] S43: Transmit the sensitive data portion to the first generator of the first transmission end, generate desensitized data and add it to the sensitive data portion in a distributed manner, and integrate and determine the pre-transmission operation data.

[0056] Specifically, according to the communication topology, the first transmission end and the second receiving end of the returned subway operation data are determined. In the data transmission process, the first transmission end is the node that sends data, and the second receiving end is the node that receives data. The two nodes can be the same or different subway data ports, or a collection of multiple data ports.

[0057] Scan the subway operation data that needs to be transmitted back, identify the data classification identifier, and then determine the sensitive data part contained in the transmitted data. Transmit the sensitive data part to the first generator of the first transmission end. The first generator processes the sensitive data part to generate desensitized data. Then add the generated desensitized data to the sensitive data part in a distributed manner. This distributed addition may be performed according to certain rules or data structures. For example, if the sensitive data part is a record containing multiple fields, the desensitized data can be added to the corresponding position according to the field correspondence. Finally, integrate the sensitive data part and the desensitized data to generate pre-transmission operation data, which will be transmitted to the second receiving end to ensure data privacy and security during data sharing.

[0058] Furthermore, step S4 of the embodiment of the present application also includes: S44: Determine a communication channel according to the transmission requirements and the channel status in combination with the communication topology.

[0059] S45: Based on the communication channel, the pre-transmission operation data is transmitted to the second receiving end, and based on the second discriminator of the second receiving end, data judgment and restoration are performed on the pre-transmission operation data.

[0060] Specifically, the current channel status of the communication topology is analyzed, and the best communication channel is determined according to the data transmission requirements. For example, if the amount of data to be transmitted is large and the channel status is good, a wired channel with a large bandwidth can be selected; if the amount of data is small and the channel status is average, a wireless channel with a small bandwidth can be selected.

[0061] Based on the determined communication channel, the pre-transmission operation data is transmitted to the second receiving end. During the transmission process, various communication protocols (such as TCP / IP protocol, FTP protocol) can be used to ensure the reliable transmission of data. After the data arrives at the second receiving end, the discriminator analyzes the received data to determine whether it meets the preset authenticity and integrity standards. If the discriminator considers the data qualified, the desensitized data is restored to the original data. For example, if the desensitized passenger information is transmitted, the second discriminator will check whether the information is complete and has not been tampered with, and then restore it to the original passenger name and contact information.

[0062] The above steps ensure the security and effectiveness of data during transmission, enabling subway operation data to be shared efficiently and reliably among different systems.

[0063] In summary, the subway operation data integration and sharing method provided in the embodiment of the present application has the following technical effects: Obtain distributed subway data ports, combine with the transmission network for interactive operation data, generate communication topology, and build a basic network architecture for subsequent data processing, interactive sharing and other operations. The data type of interactive subway operation is classified according to the preset sensitivity and the data classification mark is determined, so that the subsequent data processing can be targeted according to the different properties (sensitivities) of the data, laying the foundation for building a suitable data processing model. Based on the data classification mark, a data processing model is constructed based on an adversarial network, including a first generator and a second discriminator. This model structure can effectively process subway operation data. The data processing model is decoupled and arranged on the subway data port based on the communication topology, so as to utilize the previously established communication topology structure to perform effective data processing on each data port. The subway operation data is transmitted back. During the data transmission process, the data is desensitized based on the first generator at the data transmission end to protect data security; the data is restored based on the second discriminator at the data receiving end to ensure data availability. Combined with the communication topology, data interactive sharing management is performed to achieve secure sharing of subway operation data.

[0064] Overall, the embodiment of the present application builds a complete subway operation data integration and sharing system, from the construction of basic communication topology, to the detailed classification of data, to the construction and layout of the core data processing model, and finally to the interactive sharing management of data. This solution effectively improves the security of subway operation data, protects sensitive data through data desensitization and restoration mechanisms, improves the efficiency and accuracy of data management, and enables different types of data to be properly processed; at the same time, it also enhances the effectiveness of data interactive sharing, and provides accurate and timely data support for all aspects of subway operations.

[0065] Embodiment 2, as Figure 2 As shown, the embodiment of the present application provides a subway operation data integration and sharing system, the system comprising: The communication topology generation module 10 is used to obtain distributed subway data ports and generate communication topology in combination with the transmission network for operation data interaction. The subway data port is a multi-line data terminal connected to the subway operation information platform.

[0066] The data classification module 20 is used to interact with the data type of subway operation, perform data binary classification, and determine the data classification identifier, wherein the binary classification standard is a preset sensitivity.

[0067] A data processing model construction module 30 is used to construct a data processing model with the data classification identifier, decouple the data processing model, and deploy it at the subway data port based on the communication topology; wherein, constructing the data processing model includes: identifying the data classification identifier, determining the true sample data based on the sensitive data type, and training the first generator and the second discriminator based on the alternating iteration of the adversarial network; wherein, the first generator and the second discriminator are initialized with parameter assignment, and a random noise vector is introduced as the input of the first generator to determine the generated data once; in combination with the second discriminator, the generated data and the true sample data are determined to determine the first judgment data; based on the first judgment data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met to obtain the data processing model; wherein the adversarial equilibrium condition is that the generated data of the first generator meets the preset authenticity, and the second discriminator meets the preset judgment probability for the generated data of the first generator; the splitting of the first generator and the second discriminator is used as the decoupling method.

[0068] The data interaction and sharing management module 40 is used to transmit subway operation data back, generate obfuscated data based on the first generator at the data transmission end, insert obfuscation and desensitize sensitive data, transmit it to the data receiving end, and perform obfuscation discrimination and data restoration based on the second discriminator.

[0069] Furthermore, the data processing model building module 30 of the embodiment of the present application is also used to perform the following steps: Construct a discriminant loss function and a generation loss function, wherein the discriminant loss function is the cross entropy loss of the upper generated sample and the true sample data, and the generation loss function is the negative log-likelihood of the true probability of the upper discriminant result; based on the discriminant loss function, adjust the parameter assignment of the second discriminator after one training, and determine the secondary parameter assignment, wherein minimizing the loss is the adjustment direction; temporarily fix the secondary parameter assignment, and perform parameter assignment adjustment and secondary generation on the first generator.

[0070] Furthermore, the data processing model building module 30 of the embodiment of the present application is also used to perform the following steps: The data processing model is decoupled to split the first generator and the second discriminator; the communication topology is traversed to determine the data port type of each subway data port based on data interaction characteristics, wherein the data port type includes an output port, a receiving port and a bidirectional port; based on the data port type, the first generator and the second discriminator are configured with ports.

[0071] Furthermore, the data interaction sharing management module 40 of the embodiment of the present application is also used to perform the following steps: For the returned subway operation data, a first transmission end and a second receiving end are determined, wherein the first transmission end and the second receiving end include at least one subway data port; based on the data classification identifier, the returned subway operation data is scanned to identify the sensitive data part; the sensitive data part is transmitted to the first generator of the first transmission end, desensitized data is generated and distributedly added to the sensitive data part, and pre-transmission operation data is integrated and determined.

[0072] Furthermore, the data interaction sharing management module 40 of the embodiment of the present application is also used to perform the following steps: In combination with the communication topology, the communication channel is determined according to the transmission requirements and the channel status; based on the communication channel, the pre-transmission operation data is transmitted to the second receiving end, and based on the second discriminator of the second receiving end, the pre-transmission operation data is judged and restored.

[0073] Through the above detailed description of the subway operation data integration and sharing method in this specification, those skilled in the art can clearly understand the subway operation data integration and sharing system in this embodiment. For the system disclosed in Example 2, since it corresponds to the method disclosed in Example 1 and has corresponding functional modules and beneficial effects, the relevant parts can be referred to the method part description.

[0074] Embodiment three, based on the same inventive concept of the subway operation data integration and sharing method in the aforementioned embodiment one, the present application also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the steps of any one of the methods described in the aforementioned embodiment one.

[0075] like Figure 3 As shown, the bus architecture is represented by bus 300, which may include any number of interconnected buses and bridges, and bus 300 connects various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also connect various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, namely a transceiver, providing a unit for communicating with various other devices on a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 may be used to store data used by processor 302 when performing operations.

[0076] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A subway operation data integration and sharing method, characterized in that: The method comprises: Obtain distributed subway data ports, and generate communication topology in combination with the transmission network for interactive operation data. The subway data ports are multi-line data terminals connected to the subway operation information platform. The data type of interactive subway operation is classified into two categories to determine the data classification identifier, wherein the classification standard is the preset sensitivity; Using the data classification identifier, a data processing model is constructed, the data processing model is decoupled, and is deployed on a subway data port based on the communication topology; Among them, building a data processing model includes: identifying data classification identifiers, determining true sample data based on sensitive data types, and training a first generator and a second discriminator based on alternating iterations of an adversarial network; Among them, the parameters of the first generator and the second discriminator are initialized, and a random noise vector is introduced as the input of the first generator to determine the generated data once; the generated data and the true sample data are combined with the second discriminator to determine the first judgment data; based on the first judgment data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met to obtain the data processing model; The adversarial equilibrium condition is that the data generated by the first generator meets the preset authenticity, and the second discriminator meets the preset discrimination probability for the data generated by the first generator; the decoupling method is to split the first generator and the second discriminator; The subway operation data is transmitted back, and the first generator at the data transmission end generates obfuscated data to insert obfuscation and desensitize the sensitive data, and then transmits it to the data receiving end, and the obfuscation judgment and data restoration are performed based on the second discriminator.

2. The subway operation data integration and sharing method according to claim 1, characterized in that: Based on the primary judgment data, performing loss compensation for parameter assignment on the second discriminator and performing secondary training on the first generator include: Constructing a discriminant loss function and a generative loss function, wherein the discriminant loss function is the cross entropy loss of the upper generated sample and the true sample data, and the generative loss function is the negative log-likelihood of the true probability of the upper discriminant result; Based on the discriminant loss function, adjusting the parameter assignment of the second discriminator after the first training, and determining the secondary parameter assignment, wherein minimizing the loss is the adjustment direction; The secondary parameter assignment is temporarily fixed, and the parameter assignment adjustment and secondary generation are performed on the first generator.

3. The subway operation data integration and sharing method according to claim 1, characterized in that: The communication topology is deployed on the subway data port, including: Decoupling the data processing model, splitting the first generator and the second discriminator; Traversing the communication topology, and determining the data port type of each subway data port based on the data interaction characteristics, wherein the data port type includes an output port, a receiving port, and a bidirectional port; Based on the data port type, ports of the first generator and the second discriminator are configured.

4. The subway operation data integration and sharing method according to claim 1, characterized in that: The first generator based on the data transmission end performs data desensitization, including: For the returned subway operation data, determine a first transmission end and a second receiving end, wherein the first transmission end and the second receiving end include at least one subway data port; Based on the data classification identifier, the returned subway operation data is scanned to identify sensitive data parts; The sensitive data portion is transmitted to the first generator of the first transmission end, desensitized data is generated and added to the sensitive data portion in a distributed manner, and pre-transmission operation data is integrated and determined.

5. The subway operation data integration and sharing method according to claim 4, characterized in that: The method of restoring data based on the second discriminator at the data receiving end includes: In combination with the communication topology, determining a communication channel according to transmission requirements and channel status; Based on the communication channel, the pre-transmission operation data is transmitted to the second receiving end, and based on the second discriminator of the second receiving end, data determination and restoration are performed on the pre-transmission operation data.

6. The subway operation data integration and sharing system is characterized by: The system is used to execute the subway operation data integration and sharing method according to any one of claims 1 to 5, comprising: A communication topology generation module, which is used to obtain distributed subway data ports and generate a communication topology in combination with a transmission network for interactive operation data. The subway data ports are multi-line data terminals that are connected to a subway operation information platform. A data classification module, which is used to interact with the data type of subway operation, perform data binary classification, and determine the data classification identifier, wherein the binary classification standard is a preset sensitivity; A data processing model construction module, the data processing model construction module is used to construct a data processing model based on the data classification identifier, decouple the data processing model, and deploy it at the subway data port based on the communication topology; Among them, building a data processing model includes: identifying data classification identifiers, determining true sample data based on sensitive data types, and training a first generator and a second discriminator based on alternating iterations of an adversarial network; Among them, the first generator and the second discriminator are initialized with parameter assignment, a random noise vector is introduced as the input of the first generator, and the generated data is determined once; the second discriminator is combined, and the generated data and the true sample data are judged to determine the first judgment data; based on the first judgment data, the loss of parameter assignment is compensated for the second discriminator, and the first generator is trained twice until the adversarial equilibrium condition is met, and the data processing model is obtained; The adversarial equilibrium condition is that the data generated by the first generator meets the preset authenticity, and the second discriminator meets the preset discrimination probability for the data generated by the first generator; the decoupling method is to split the first generator and the second discriminator; A data interaction and sharing management module is used to transmit subway operation data back, generate obfuscated data based on the first generator at the data transmission end, insert obfuscation and desensitize sensitive data, transmit it to the data receiving end, and perform obfuscation discrimination and data restoration based on the second discriminator.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the subway operation data integration and sharing method described in any one of claims 1-5 are implemented.

Images