A data processing method, device, terminal, and medium based on data desensitization

Abstract

The present application discloses a data processing method, apparatus, terminal, and medium based on data desensitization. The method includes: retrieving at least two original data files from a data file database and obtaining the confidentiality level labels corresponding to the original data files; performing desensitization processing on the original data files based on the confidentiality level labels to obtain processed original data files; obtaining the confidentiality evaluation results corresponding to the processed original data files according to the processed original data files; determining the desensitization deviation values corresponding to the original data files based on the confidentiality level labels and the confidentiality evaluation results; if there are desensitization deviation values not less than a first preset ratio that are less than a preset desensitization deviation threshold, determining the processed original data files as target data files. The purpose of the present application is to ensure that the data desensitization processing effect of the original data files in the same batch reaches the expected level and improve the confidentiality management level of the data files.

Description

Technical Field

[0001] This application relates to the technical field of data desensitization, and in particular, to a data processing method, device, terminal, and medium based on data desensitization. Background Art

[0002] Data Masking is a technology that protects privacy by modifying sensitive data. It ensures that, while maintaining the original format and statistical characteristics of the data, personal or other sensitive information is not leaked.

[0003] However, for data files with different confidentiality levels in a batch or a database, the existing data desensitization methods have the following deficiencies: the existing data confidentiality management strategies cannot perform differential processing on data files with different confidentiality levels. This may result in insufficient protection for data files with a high confidentiality level, while data files with a low confidentiality level may be over-desensitized, affecting the usability of the data; there is a lack of quantitative evaluation indicators, which leads to strong subjectivity in evaluation results and makes it difficult to conduct objective comparison and verification; the existing desensitization methods usually do not establish an effective feedback mechanism, unable to timely detect and correct problems in the desensitization process, resulting in unstable desensitization effects and thus unable to overall improve the data confidentiality management level. Summary of the Invention

[0004] The main purpose of this application is to provide a data processing method, device, terminal, and medium based on data desensitization, aiming to ensure that the data desensitization processing effect of the original data files in the same batch reaches the expected level, enabling files with different confidentiality levels to adapt to different data confidentiality effects, and continuously enhancing the data desensitization effect by quantifying the error of the desensitization process, thereby improving the overall confidentiality management level of the data files.

[0005] To achieve the above objective, this application provides a data processing method based on data desensitization, which is applied to a data middle platform. The data middle platform includes a data file database and a server, and both the data file database and the server are in the same network;

[0006] The method includes:

[0007] Retrieve at least two original data files from the data file database, and obtain the confidentiality level label corresponding to the original data file. Herein, the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels;

[0008] Based on the confidentiality level label, perform desensitization processing on the original data file to obtain the processed original data file;

[0009] By presetting a confidentiality evaluation model, according to the processed original data file, obtain a confidentiality evaluation result corresponding to the processed original data file;

[0010] Based on the confidentiality level label and the confidentiality evaluation result, determine a desensitization deviation value corresponding to the original data file;

[0011] If there are desensitization deviation values not less than a first preset ratio among the desensitization deviation values that are less than a preset desensitization deviation threshold, then determine the processed original data file as the target data file.

[0012] Specifically, a preset desensitization processing model set is loaded in the server, the preset desensitization processing model set includes at least two different preset desensitization processing models, and the preset desensitization processing models correspond to the confidentiality levels one by one;

[0013] The desensitizing the original data file based on the confidentiality level label to obtain a processed original data file includes:

[0014] Based on the confidentiality level label, select a preset desensitization processing model corresponding to the confidentiality level label from the preset desensitization processing model set;

[0015] Through the preset desensitization processing model corresponding to the confidentiality level label, according to the original data file, obtain the processed original data file.

[0016] Specifically, the preset desensitization processing model includes a first input layer, a first embedding layer, a first convolutional layer, a pooling layer, a first fully connected layer, and a first output layer;

[0017] The obtaining the processed original data file through the preset desensitization processing model corresponding to the confidentiality level label according to the original data file includes:

[0018] Through the first input layer, according to the sensitive information in the original data file, obtain a first sensitive information tensor;

[0019] Through the first embedding layer, according to the first sensitive information tensor, obtain a first floating-point tensor;

[0020] Through the first convolutional layer, according to the first floating-point tensor, obtain a second floating-point tensor;

[0021] Through the pooling layer, according to the second floating-point tensor, obtain a third floating-point tensor;

[0022] Through the first fully connected layer, according to the third floating-point tensor, obtain a fourth floating-point tensor;

[0023] Through the first output layer, a first output tensor is obtained according to the fourth floating-point tensor, where the first output tensor is used to represent the desensitized sensitive information;

[0024] Based on the original data file and the first output tensor, the processed original data file is obtained.

[0025] Specifically, the preset desensitization processing model includes a second input layer, a second embedding layer, a second convolutional layer, a recurrent neural network layer, an attention mechanism layer, a second fully-connected layer, and a second output layer;

[0026] The obtaining of the processed original data file according to the original data file through the preset desensitization processing model corresponding to the confidentiality level label includes:

[0027] Through the second input layer, a second sensitive information tensor is obtained according to the sensitive information in the original data file;

[0028] Through the second embedding layer, a fifth floating-point tensor is obtained according to the second sensitive information tensor;

[0029] Through the second convolutional layer, a sixth floating-point tensor is obtained according to the fifth floating-point tensor;

[0030] Through the recurrent neural network layer, a seventh floating-point tensor is obtained according to the sixth floating-point tensor;

[0031] Through the attention mechanism layer, an eighth floating-point tensor is obtained according to the seventh floating-point tensor;

[0032] Through the second fully-connected layer, a ninth floating-point tensor is obtained according to the eighth floating-point tensor;

[0033] Through the second output layer, a second output tensor is obtained according to the ninth floating-point tensor, where the second output tensor is used to represent the desensitized sensitive information;

[0034] Based on the original data file and the second output tensor, the processed original data file is obtained.

[0035] Specifically, the preset confidentiality evaluation model includes a third input layer, a third embedding layer, a self-attention mechanism layer, a multi-layer perceptron layer, and a third output layer;

[0036] The obtaining of the confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through the preset confidentiality evaluation model includes:

[0037] Through the third input layer, an input tensor is obtained according to the processed original data file;

[0038] Through the third embedding layer, a tenth floating-point tensor is obtained according to the input tensor;

[0039] Through the self-attention mechanism layer, an eleventh floating-point tensor is obtained according to the tenth floating-point tensor;

[0040] Through the multi-layer perceptron layer, a twelfth floating-point tensor is obtained according to the eleventh floating-point tensor;

[0041] Through the third output layer, a confidentiality evaluation score is obtained according to the twelfth floating-point tensor, and the confidentiality evaluation score is determined as the confidentiality evaluation result.

[0042] Specifically, determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result includes:

[0043] Converting the confidentiality level corresponding to the confidentiality level label into a confidentiality level value corresponding to the confidentiality level label through a preset linear mapping function;

[0044] Calculating the absolute value of the difference between the confidentiality level value and the confidentiality evaluation score to obtain the desensitization deviation value.

[0045] Specifically, after determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result, the method further includes:

[0046] If there are desensitization deviation values not less than a second preset ratio among the desensitization deviation values that are greater than a preset desensitization deviation threshold, then based on the desensitization deviation values, update the preset desensitization processing model to obtain an updated preset desensitization processing model, where the updated preset desensitization model is used to perform desensitization processing on the original data file.

[0047] To achieve the above object, the present application further provides a data processing device based on data desensitization, which is applied to a data middle platform. The data middle platform includes a data file database and a server, and both the data file database and the server are in the same network;

[0048] The device includes:

[0049] A first unit, configured to retrieve at least two original data files from the data file database and obtain the confidentiality level label corresponding to the original data file, where the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels;

[0050] A second unit, configured to desensitize the original data file based on the confidentiality level label to obtain a processed original data file;

[0051] A third unit, configured to obtain a confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through a preset confidentiality evaluation model;

[0052] A fourth unit, configured to determine a desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result;

[0053] A fifth unit, configured to determine the processed original data file as a target data file if there is a desensitization deviation value not less than a first preset ratio of the desensitization deviation values that is less than a preset desensitization deviation threshold.

[0054] To achieve the above object, the present application further provides a terminal, including a memory storing multiple instructions; the processor loads the instructions from the memory to execute the steps in any one of the methods provided by the present application.

[0055] To achieve the above object, the present application further provides a computer storage medium storing multiple instructions, and the instructions are suitable for being loaded by a processor to execute the steps in any one of the methods provided by the present application.

[0056] A data processing method, apparatus, terminal and medium based on data desensitization provided by the present application can first retrieve at least two original data files from a data file database and obtain confidentiality level labels corresponding to the original data files, where the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels; then, desensitize the original data file based on the confidentiality level label to obtain a processed original data file; then, obtain a confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through a preset confidentiality evaluation model; then, determine a desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result; finally, if there is a desensitization deviation value not less than a first preset ratio of the desensitization deviation values that is less than a preset desensitization deviation threshold, determine the processed original data file as a target data file to ensure that the data desensitization processing effect of the original data files in the same batch reaches the expected level.

[0057] Compared with the existing data desensitization methods, the present application can ensure that the data desensitization effect of the original data files in the same batch reaches the expected level, enabling files with different confidentiality levels to adapt to different data confidentiality effects. By quantifying the error of the desensitization process, the data desensitization effect is continuously enhanced, thereby improving the overall confidentiality management level of the data files. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic flowchart provided by an embodiment of the present application;

[0059] Figure 2 It is a schematic structural diagram of a data processing device based on data desensitization provided by an embodiment of the present application;

[0060] Figure 3 It is a schematic structural diagram of a terminal provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0062] Since for data files with different confidentiality levels in a batch or a database, the existing data desensitization methods have the following deficiencies: the existing data confidentiality management strategies cannot perform differential processing on data files with different confidentiality levels. This may result in insufficient protection for data files with high confidentiality levels, while data files with low confidentiality levels may be over-desensitized, affecting the availability of the data; there is a lack of quantitative evaluation indicators, which leads to strong subjectivity in the evaluation results and makes it difficult to conduct objective comparison and verification; the existing desensitization methods usually do not establish an effective feedback mechanism, and cannot timely discover and correct problems in the desensitization process, resulting in unstable desensitization effects and thus unable to improve the overall data confidentiality management level.

[0063] Therefore, the embodiments of the present application provide a data processing method, device, terminal, and medium based on data desensitization to solve practical technical problems.

[0064] In some embodiments, the device may be specifically integrated in an electronic device, and the electronic device may be a device such as a terminal or a server.

[0065] In some embodiments, the server may also be implemented in the form of a terminal.

[0066] Among them, the server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms.

[0067] Among them, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server can be directly or indirectly connected through wired or wireless communication methods, and this application does not make any restrictions here.

[0068] The following will be described in detail respectively. It should be noted that the serial numbers of the following embodiments do not limit the preferred order of the embodiments.

[0069] The embodiment of the present application provides a data processing method based on data desensitization. The method can ensure that the data desensitization processing effect of the original data files in the same batch reaches the expected level, enable files with different confidentiality levels to adapt to different data confidentiality effects, and continuously enhance the data desensitization effect by quantifying the error of the desensitization processing, thereby improving the overall confidentiality management level of the data files.

[0070] A data processing method based on data desensitization is applied to a data middle platform. The data middle platform includes a data file database and a server, and both the data file database and the server are in the same network.

[0071] The data middle platform (Data Middleware Platform) is a type of enterprise-level data management and application architecture model, playing a core role in the modern data-driven business environment. The goal of the data middle platform is to unify and optimize the management and use of data resources to support more efficient and flexible data analysis and business decision-making. Specifically, the data middle platform has the following characteristics:

[0072] Data integration: The data middle platform integrates and standardizes data from different sources and different formats to form a unified data view, eliminating information silos and ensuring data consistency and accuracy;

[0073] Data governance: It provides a complete set of data management processes, including data quality control, data security, data life cycle management, etc., to ensure data compliance and availability;

[0074] Data analysis and services: The data middle platform supports real-time analysis, predictive analysis, and intelligent decision-making by providing preprocessed data, models, and algorithms, helping enterprises quickly respond to market changes;

[0075] Data sharing and reuse: Facilitates cross - departmental and cross - system data sharing, reduces duplicate construction, increases the data reuse rate, and lowers the costs of data acquisition and processing;

[0076] Flexibility and scalability: The data middle platform adopts a microservices architecture and can be easily extended and adjusted to adapt to the rapid development and changes of the business;

[0077] Technology stack: The data middle platform usually involves components such as big data processing technologies, data warehouses, data lakes, data APIs, machine learning platforms, etc.;

[0078] In some embodiments, the data middle platform includes a data file repository and a server. Both the data file repository and the server are in the same network. The data file repository can be used to store different original data files corresponding to different confidentiality level tags, and the server is used to manage, process the original data files and the original data files after desensitization processing.

[0079] As Figure 1 shown, the specific process of the method can be as follows:

[0080] S110. Retrieve at least two original data files from the data file repository and obtain the confidentiality level tags corresponding to the original data files. Among them, the confidentiality level tag is used to represent one selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels.

[0081] In some embodiments, in the data file repository, the batch identifier to which each original data file belongs can be recorded, and a unique batch identifier (Batch ID) can be assigned to the original data files of each batch. This identifier can be a string or a number and is used to uniquely identify a batch of original data files; or when retrieving data files, the batch identifier can be specified to ensure that the retrieved original data files belong to the same batch.

[0082] In some embodiments, a preset confidentiality level set can be predefined, that is, a preset confidentiality level set is determined, for example, {'Low', 'High'}, which represents the confidentiality level of the original data file. At the same time, in the data file repository, the corresponding confidentiality level tag of each original data file can be stored, usually implemented through a database table or other data structures.

[0083] Specifically, the corresponding confidentiality level tag of each original data file can be determined in the following ways:

[0084] Manual annotation:

[0085] The confidentiality level is manually marked by data security experts or business personnel according to the content and sensitivity of the data file. This method is applicable to situations with a small amount of data and high accuracy requirements; the confidentiality level of the data file is manually marked according to existing business rules and standards. For example, financial data files may be default marked as "High", while user basic information files may be marked as "Low";

[0086] Automated marking:

[0087] Natural Language Processing (NLP): Use NLP technology to automatically analyze the content of the data file, identify sensitive information therein, and automatically mark the confidentiality level according to preset rules; Keyword matching: Automatically mark the confidentiality level by matching a preset list of sensitive keywords. For example, files containing keywords such as "bank account", "password", etc. can be marked as "High"; Entity recognition: Use Named Entity Recognition (NER) technology to identify sensitive entities (such as names, addresses, phone numbers, etc.) in the file, and automatically mark the confidentiality level according to the recognition results; Machine learning model: Train a machine learning model to automatically predict the confidentiality level of new data files based on the marking results of historical data files; Supervised learning: Use labeled data files to train classification models (such as logistic regression, decision tree, random forest, etc.), and then use the model to predict the confidentiality level of new data files; Unsupervised learning: Use clustering algorithms (such as K-means, DBSCAN, etc.) to cluster data files, and manually mark or further analyze according to the clustering results;

[0088] Metadata marking:

[0089] File metadata: Utilize the metadata of the file (such as file type, creation date, modification date, etc.) to assist in determining the confidentiality level; for example, financial report files may be default marked as "High"; File path: Determine the confidentiality level according to the storage path of the file; for example, files stored in a specific secure directory can be default marked as "High";

[0090] Rule engine:

[0091] Rule definition: Define a set of rules to automatically mark the confidentiality level according to the content and attributes of the file. The rules can include multiple conditions such as keyword matching, file path, file type, etc.; Example rules: If the file contains the keyword "bank account", it is marked as "High"; If the file is stored in the "financial data" directory, it is marked as "High"; If the file creation date is within the last 30 days, it is marked as "High".

[0092] S120. Desensitize the original data file based on the confidentiality level label to obtain a processed original data file.

[0093] In some embodiments, a preset desensitization processing model set is installed in the server. The preset desensitization processing model set includes at least two different preset desensitization processing models, and the preset desensitization processing models correspond to the confidentiality levels one by one.

[0094] Specifically, the step of desensitizing the original data file based on the confidentiality level label to obtain a processed original data file includes the following steps S121 to S122:

[0095] S121. Based on the confidentiality level label, select the preset desensitization processing model corresponding to the confidentiality level label from the preset desensitization processing model set.

[0096] S122. Through the preset desensitization processing model corresponding to the confidentiality level label, obtain the processed original data file according to the original data file.

[0097] In some embodiments, the preset desensitization processing model includes a first input layer, a first embedding layer, a first convolutional layer, a pooling layer, a first fully connected layer, and a first output layer.

[0098] In some embodiments, the preset desensitization processing model including a first input layer, a first embedding layer, a first convolutional layer, a pooling layer, a first fully connected layer, and a first output layer can be a neural network model applicable to the original data file with a low confidentiality level. The neural network model applicable to the original data file with a low confidentiality level is used for scenarios with relatively low data confidentiality requirements, such as internal testing and development environments. This model structure is relatively simple, with a fast processing speed, but the desensitization intensity is slightly lower.

[0099] Specifically, the step of obtaining the processed original data file through the preset desensitization processing model corresponding to the confidentiality level label according to the original data file includes the following steps A1 to A7:

[0100] A1. Through the first input layer, obtain a first sensitive information tensor according to the sensitive information in the original data file.

[0101] In some embodiments, through the first input layer, the sensitive information in the original data file can be converted into a two-dimensional tensor with a shape of (batch_size, sequence_length), where batch_size is the size of the batch of sensitive information, and sequence_length is the length of the information sequence in the sensitive information. This two-dimensional tensor is the first sensitive information tensor.

[0102] A2. Through the first embedding layer, a first floating-point tensor is obtained according to the first sensitive information tensor.

[0103] In some embodiments, through the first embedding layer, each word in the text data of the sensitive information is converted into a vector representation of a fixed length, that is, the discrete vocabulary index is converted into a representation in a continuous vector space, which helps subsequent feature extraction.

[0104] Specifically, the first floating-point tensor is a floating-point tensor with a shape of (batch_size, sequence_length, embedding_dim), where embedding_dim is the dimension of the embedding vector.

[0105] A3. Through the first convolutional layer, a second floating-point tensor is obtained according to the first floating-point tensor.

[0106] In some embodiments, the first convolutional layer can extract local features using multiple convolution kernels of different sizes. Each convolution kernel slides on the input data to generate a feature map; multiple convolution kernels can capture local features of different scales, and then form a floating-point tensor.

[0107] Specifically, through the first convolutional layer, according to the first floating-point tensor, a floating-point tensor with a shape of (batch_size, sequence_length - kernel_size + 1, num_filters) is obtained, that is, the second floating-point tensor, where kernel_size is the size of the convolution kernel, and num_filters is the number of convolution kernels.

[0108] A4. Through the pooling layer, a third floating-point tensor is obtained according to the second floating-point tensor.

[0109] In some embodiments, the pooling layer reduces the feature dimension through max pooling or average pooling, retains the most important information, can reduce the complexity of the model, and reduce the risk of overfitting.

[0110] Specifically, through the pooling layer, a floating-point tensor with a shape of (batch_size, (sequence_length - kernel_size + 1) / / pool_size, num_filters) is obtained according to the second floating-point tensor, that is, the third floating-point tensor, where pool_size is the size of the pooling window.

[0111] A5. Through the first fully connected layer, a fourth floating-point tensor is obtained according to the third floating-point tensor.

[0112] In some embodiments, the first fully connected layer performs feature fusion and classification on the feature vectors after convolution and pooling, and can capture global features to promote the generation of the final desensitization result.

[0113] Specifically, through the first fully connected layer, a floating-point tensor with a shape of (batch_size, output_dim) is obtained according to the third floating-point tensor, that is, the fourth floating-point tensor, where output_dim is the output dimension of the first fully connected layer.

[0114] A6. Through the first output layer, a first output tensor is obtained according to the fourth floating-point tensor, where the first output tensor is used to represent the sensitive information after desensitization processing.

[0115] In some embodiments, the fourth floating-point tensor is input into the first output layer, and an integer or string tensor with a shape of (batch_size, sequence_length) is output, that is, the first output tensor, which is used to represent the sensitive information after desensitization processing. For example, it can be the asterisk part that replaces the middle part of the phone number.

[0116] A7. Based on the original data file and the first output tensor, the processed original data file is obtained.

[0117] In some embodiments, the sensitive information part in the original data file can be replaced with the desensitized sensitive information to obtain the processed original data file.

[0118] In some embodiments, the preset desensitization processing model includes a second input layer, a second embedding layer, a second convolutional layer, a recurrent neural network layer, an attention mechanism layer, a second fully connected layer, and a second output layer.

[0119] In some embodiments, a preset desensitization processing model including a second input layer, a second embedding layer, a second convolutional layer, a recurrent neural network layer, an attention mechanism layer, a second fully connected layer, and a second output layer can be a neural network model applicable to original data files with a high security level. The neural network model applicable to original data files with a high security level is used for scenarios with high data confidentiality requirements, such as external data sharing and compliance auditing. This model has a complex structure, a relatively slow processing speed, but a high desensitization intensity and stronger security.

[0120] Specifically, obtaining the processed original data file according to the original data file through the preset desensitization processing model corresponding to the security level label includes the following steps B1 to B8:

[0121] B1. Through the second input layer, according to the sensitive information in the original data file, obtain a second sensitive information tensor.

[0122] In some embodiments, through the second input layer, the sensitive information in the original data file can be converted into a two-dimensional tensor with a shape of (batch_size, sequence_length), where batch_size is the size of the batch of sensitive information, and sequence_length is the length of the information sequence in the sensitive information. This two-dimensional tensor is the second sensitive information tensor.

[0123] B2. Through the second embedding layer, according to the second sensitive information tensor, obtain a fifth floating-point tensor.

[0124] In some embodiments, through the second embedding layer, each word in the text data of the sensitive information is converted into a fixed-length vector representation, that is, the discrete vocabulary index is converted into a representation in a continuous vector space, which helps subsequent feature extraction.

[0125] Specifically, the fifth floating-point tensor is a floating-point tensor with a shape of (batch_size, sequence_length, embedding_dim), where embedding_dim is the dimension of the embedding vector.

[0126] B3. Through the second convolutional layer, according to the fifth floating-point tensor, obtain a sixth floating-point tensor.

[0127] In some embodiments, the second convolutional layer can use multiple convolution kernels of different sizes to extract local features. Each convolution kernel slides on the input data to generate a feature map; multiple convolution kernels can capture local features of different scales, and then form a floating-point tensor.

[0128] Specifically, through the second convolutional layer, a floating-point tensor of shape (batch_size, sequence_length - kernel_size + 1, num_filters) is obtained according to the fifth floating-point tensor, that is, the sixth floating-point tensor, where kernel_size is the size of the convolutional kernel and num_filters is the number of convolutional kernels.

[0129] B4. Through the recurrent neural network layer, a seventh floating-point tensor is obtained according to the sixth floating-point tensor.

[0130] In some embodiments, the recurrent neural network layer uses a bidirectional LSTM layer to capture long-range dependencies. The bidirectional LSTM can capture sequence information from both the forward and backward directions to improve the expressive ability of the model.

[0131] Specifically, through the recurrent neural network layer, a floating-point tensor of shape (batch_size, sequence_length - kernel_size + 1, hidden_dim * 2) is obtained according to the sixth floating-point tensor, that is, the seventh floating-point tensor, where hidden_dim is the hidden state dimension of the LSTM cell and * 2 represents the bidirectional LSTM.

[0132] B5. Through the attention mechanism layer, an eighth floating-point tensor is obtained according to the seventh floating-point tensor.

[0133] In some embodiments, the attention mechanism layer focuses on key information by calculating the importance weights at each time step to improve the accuracy of the model and thus enhance the effect of the desensitization process.

[0134] Specifically, through the attention mechanism layer, a floating-point tensor of shape (batch_size, hidden_dim * 2) is obtained according to the seventh floating-point tensor, that is, the eighth floating-point tensor.

[0135] B6. Through the second fully connected layer, a ninth floating-point tensor is obtained according to the eighth floating-point tensor.

[0136] In some embodiments, the second fully connected layer performs feature fusion and classification on the feature vectors after the attention mechanism, and can capture global features to promote the generation of the final desensitization result.

[0137] Specifically, through the second fully-connected layer, a floating-point tensor with a shape of (batch_size, output_dim) is obtained according to the eighth floating-point tensor, that is, the ninth floating-point tensor, where output_dim is the output dimension of the fully-connected layer.

[0138] B7. Through the second output layer, a second output tensor is obtained according to the ninth floating-point tensor, where the second output tensor is used to represent the desensitized sensitive information.

[0139] In some embodiments, the ninth floating-point tensor is input into the second output layer, and an integer or string tensor with a shape of (batch_size, sequence_length) is output, that is, the second output tensor, which is used to represent the desensitized sensitive information. For example, it can be the random number part that replaces the middle part of the phone number.

[0140] B8. Based on the original data file and the second output tensor, the processed original data file is obtained.

[0141] In some embodiments, the part where the sensitive information is located in the original data file can be replaced with the desensitized sensitive information to obtain the processed original data file.

[0142] S130. Through a preset confidentiality evaluation model, according to the processed original data file, a confidentiality evaluation result corresponding to the processed original data file is obtained.

[0143] In some embodiments, the preset confidentiality evaluation model includes a third input layer, a third embedding layer, a self-attention mechanism layer, a multi-layer perceptron layer, and a third output layer.

[0144] Specifically, the process of obtaining the confidentiality evaluation result corresponding to the processed original data file through the preset confidentiality evaluation model includes the steps S131 to S135 as follows:

[0145] S131. Through the third input layer, an input tensor is obtained according to the processed original data file.

[0146] In some embodiments, the input tensor can be a two-dimensional tensor with a shape of (batch_size, sequence_length), where batch_size is the batch size of the processed sensitive information, and sequence_length is the length of each sequence of the processed sensitive information.

[0147] S132. Obtain the tenth floating-point tensor according to the input tensor through the third embedding layer.

[0148] In some embodiments, through the third embedding layer, a floating-point tensor with a shape of (batch_size, sequence_length, embedding_dim) is obtained according to the input tensor, that is, the tenth floating-point tensor, where embedding_dim is the dimension of the embedding vector.

[0149] S133. Obtain the eleventh floating-point tensor according to the tenth floating-point tensor through the self-attention mechanism layer.

[0150] In some embodiments, the self-attention mechanism layer can capture long-range dependencies in the sequence, focus on key information by calculating the importance weights at each time step, so as to improve the accuracy of the model.

[0151] Specifically, through the third embedding layer, a floating-point tensor with a shape of (batch_size, sequence_length, attention_dim) is obtained according to the tenth floating-point tensor, that is, the eleventh floating-point tensor, where attention_dim is the output dimension of the attention mechanism.

[0152] S134. Obtain the twelfth floating-point tensor according to the eleventh floating-point tensor through the multi-layer perceptron layer.

[0153] In some embodiments, feature fusion and classification are performed through the multi-layer perceptron layer to capture global features, so as to promote the generation of the final confidentiality evaluation result.

[0154] Specifically, through the multi-layer perceptron layer, a floating-point tensor with a shape of (batch_size, hidden_dim) is obtained according to the eleventh floating-point tensor, that is, the twelfth floating-point tensor, where hidden_dim is the dimension of the hidden layer of the multi-layer perceptron layer.

[0155] S135. Obtain the confidentiality evaluation score according to the twelfth floating-point tensor through the third output layer, and determine the confidentiality evaluation score as the confidentiality evaluation result.

[0156] In some embodiments, through the third output layer, a floating-point tensor with a shape of (batch_size, 1) is obtained according to the twelfth floating-point tensor to represent the confidentiality evaluation score of each data file.

[0157] S140. Determine the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result.

[0158] In some embodiments, the step of determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result includes the step contents of S141 to S142 as follows:

[0159] S141. Convert the confidentiality level corresponding to the confidentiality level label into the confidentiality level value corresponding to the confidentiality level label through a preset linear mapping function.

[0160] S142. Calculate the absolute value of the difference between the confidentiality level value and the confidentiality evaluation score to obtain the desensitization deviation value.

[0161] In some embodiments, the confidentiality level label corresponding to the original data file may include "High" and "Low", and the confidentiality evaluation score may be a value within the range of [0, 1]. "High", that is, the high confidentiality level, and its interval range is [0.8, 1.0]; "Low", that is, the low confidentiality level, and its interval range is [0.6, 0.8].

[0162] Specifically, the preset linear mapping function includes an interval range, a coefficient, and a preset normalization value, and the coefficient can be determined by the size of the interval range.

[0163] For an original data file with a confidentiality level label of "High", if the coefficient is set to 0.2 and the preset normalization value is 0.75, the corresponding confidentiality level value can be expressed as:

[0164] Confidentiality level value = 0.8 + 0.2 × normalization value

[0165] For an original data file with a confidentiality level label of "Low", if the coefficient is set to 0.2 and the preset normalization value is 0.75, the corresponding confidentiality level value can be expressed as:

[0166] Confidentiality level value = 0.6 + 0.2 × normalization value

[0167] After obtaining the confidentiality level value, the desensitization deviation value is expressed by the following expression:

[0168] Desensitization deviation value = |Confidentiality level value - Actual confidentiality score|

[0169] Furthermore, based on the desensitization deviation value, evaluate whether the effect of the desensitization process has achieved the expected effect, that is, evaluate whether the adaptive desensitization operation is performed according to the existing confidentiality level.

[0170] S150. If there are desensitization deviation values not less than the first preset ratio that are less than the preset desensitization deviation threshold among the desensitization deviation values, determine the processed original data file as the target data file.

[0171] In some embodiments, 5 original data files were retrieved from the data file database for desensitization processing, and the corresponding desensitization deviation values were obtained through the above embodiment content, as shown in the following content:

[0172] For the first processed original data file file1.txt, its desensitization deviation value is 0.05;

[0173] For the second processed original data file file2.txt, its desensitization deviation value is 0.02;

[0174] For the third processed original data file file3.txt, its desensitization deviation value is 0.02;

[0175] For the fourth processed original data file file4.txt, its desensitization deviation value is 0.00;

[0176] For the fifth processed original data file file5.txt, its desensitization deviation value is 0.05.

[0177] Now, set the first preset ratio to 80%, and set the preset desensitization deviation threshold to 0.06. Then, the desensitization deviation values of the above processed original data files are all lower than the preset desensitization deviation threshold, and the proportion of the processed original data files with desensitization deviation values lower than the preset desensitization deviation threshold among the retrieved original data files is 100%, exceeding 80%. Therefore, determine the above processed original data file as the target data file that has successfully completed desensitization processing.

[0178] In some embodiments, after determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result, the method further includes the following specific implementation process:

[0179] If there are desensitization deviation values not less than the second preset ratio that are greater than the preset desensitization deviation threshold among the desensitization deviation values, update the preset desensitization processing model based on the desensitization deviation values to obtain an updated preset desensitization processing model, where the updated preset desensitization model is used to perform desensitization processing on the original data file.

[0180] In some embodiments, 5 original data files were retrieved from the data file database for desensitization processing, and the corresponding desensitization deviation values were obtained through the above embodiment content, as shown in the following content:

[0181] The first original data file after processing, file1.txt, has a desensitization deviation value of 0.05;

[0182] The second original data file after processing, file2.txt, has a desensitization deviation value of 0.02;

[0183] The third original data file after processing, file3.txt, has a desensitization deviation value of 0.02;

[0184] The fourth original data file after processing, file4.txt, has a desensitization deviation value of 0.00;

[0185] The fifth original data file after processing, file5.txt, has a desensitization deviation value of 0.05.

[0186] Now, set the second preset ratio to 20%, and set the preset desensitization deviation threshold to 0.04. Then, among the desensitization deviation values of the above-mentioned processed original data files, two are higher than the preset desensitization deviation threshold. Moreover, the proportion of the processed original data files with desensitization deviation values higher than the preset desensitization deviation threshold among the retrieved original data files is 40%, exceeding 20%. Therefore, next, based on the desensitization deviation value, the preset desensitization processing model can be updated to obtain an updated preset desensitization processing model, where the updated preset desensitization model is used to perform desensitization processing on the original data file.

[0187] Specifically, when it is found that the desensitization deviation values of file1.txt and file5.txt are relatively high, the reason is that some sensitive information has not been fully desensitized. The following steps can be taken to update the preset desensitization processing model:

[0188] First, analyze the deviation reasons. File1.txt and file5.txt contain some sensitive keywords, such as "bank account" and "ID number", but these keywords have not been fully desensitized. Then, adjust the desensitization strategy, increase the detection and replacement rules for sensitive keywords to ensure that these keywords are completely replaced during the desensitization process. More powerful desensitization algorithms can also be introduced, such as text desensitization algorithms based on deep learning, to improve the desensitization effect. Retrain the preset desensitization processing model using the new dataset (including file1.txt and file5.txt), optimize the hyperparameters of the model, and improve the generalization ability and stability of the model. Finally, use the updated preset desensitization model to perform re-desensitization processing on the original data file and calculate the desensitization deviation value again until it is ensured that the updated preset desensitization model can effectively reduce the desensitization deviation value to reach the expected desensitization processing level, ultimately improving the overall data confidentiality management level.

[0189] As can be seen from the above, in the embodiment of the present application, at first, at least two original data files can be retrieved from the data file database, and the confidentiality level label corresponding to the original data file can be obtained, where the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels; then, based on the confidentiality level label, the original data file is desensitized to obtain the processed original data file; then, through a preset confidentiality evaluation model, according to the processed original data file, the confidentiality evaluation result corresponding to the processed original data file is obtained; after that, based on the confidentiality level label and the confidentiality evaluation result, the desensitization deviation value corresponding to the original data file is determined; finally, if there are desensitization deviation values not less than a first preset ratio that are less than a preset desensitization deviation threshold, it is determined that the processed original data file is the target data file to ensure that the data desensitization processing effect of the same batch of original data files reaches the expected level.

[0190] In summary, compared with the existing data desensitization processing or data confidentiality management methods, the embodiment of the present application can ensure that the data desensitization processing effect of the same batch of original data files reaches the expected level, enable files with different confidentiality levels to adapt to different data confidentiality effects, and continuously enhance the data desensitization effect by quantifying the error of the desensitization processing, thereby improving the overall confidentiality management level of the data files.

[0191] To better implement the above method, the embodiment of the present application also provides a data processing device based on data desensitization. This device can be specifically integrated in an electronic device, and the electronic device can be a device such as a terminal or a server. Among them, the terminal can be a device such as a mobile phone, a tablet computer, a smart Bluetooth device, a laptop computer, or a personal computer; the server can be a single server or a server cluster composed of multiple servers.

[0192] For example, in this embodiment, taking the data processing device based on data desensitization being specifically integrated in the terminal as an example, the method of the embodiment of the present application will be described in detail.

[0193] For example, as Figure 2 shown, the data processing device based on data desensitization can include a first unit 201, a second unit 202, a third unit 203, a fourth unit 204, and a fifth unit 205, which is applied to a data middle platform. The data middle platform includes a data file database and a server, and both the data file database and the server are in the same network;

[0194] The device includes:

[0195] The first unit 201 is used to retrieve at least two original data files from a data file repository and obtain the confidentiality level labels corresponding to the original data files. The confidentiality level labels are used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels;

[0196] The second unit 202 is used to perform desensitization processing on the original data file based on the confidentiality level label to obtain a processed original data file;

[0197] The third unit 203 is used to obtain a confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through a preset confidentiality evaluation model;

[0198] The fourth unit 204 is used to determine a desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result;

[0199] The fifth unit 205 is used to determine the processed original data file as the target data file if there is a desensitization deviation value not less than a first preset ratio among the desensitization deviation values that is less than a preset desensitization deviation threshold.

[0200] In specific implementation, each of the above units can be implemented as an independent entity, or can be combined arbitrarily to be implemented as the same or several entities. For the specific implementation of each of the above units, reference can be made to the method embodiments above, which will not be elaborated here.

[0201] As can be seen from the above, the embodiments of the present application can ensure that the data desensitization processing effect of the same batch of original data files reaches the expected level, enabling files with different confidentiality levels to adapt to different data confidentiality effects, and continuously enhancing the data desensitization effect by quantifying the error of the desensitization processing, thereby improving the overall confidentiality management level of the data files.

[0202] The embodiments of the present application also provide an electronic device, which can be a device such as a terminal, a server, etc. Among them, the terminal can be a mobile phone, a tablet computer, a smart Bluetooth device, a laptop computer, a personal computer, etc.; the server can be a single server or a server cluster composed of multiple servers, etc.

[0203] In some embodiments, the data processing device based on data desensitization can also be integrated in multiple electronic devices. For example, the data processing device based on data desensitization can be integrated in multiple servers, and the data processing method based on data desensitization of the present application can be implemented by multiple servers.

[0204] In this embodiment, the electronic device in this embodiment will be described in detail taking the example that the electronic device is a terminal. For example, asFigure 3 As shown, it shows a schematic structural diagram of a terminal involved in an embodiment of the present application. Specifically:

[0205] The terminal may include a processor 301 with one or more processing cores, a memory 302 of one or more media, a power supply 303, an input module 304, a communication module 305, and other components. Those skilled in the art can understand that Figure 3 the terminal structure shown in does not constitute a limitation on the terminal, and it may include more or fewer components than shown, or combine certain components, or have different component arrangements. Among them:

[0206] The processor 301 is the control center of the terminal, connecting various parts of the entire terminal through various interfaces and lines. By running or executing software programs and / or modules stored in the memory 302, and by calling data stored in the memory 302, it executes various functions of the terminal and processes data, thereby monitoring the terminal as a whole. In some embodiments, the processor 301 may include one or more processing cores; in some embodiments, the processor 301 may integrate an application processor and a modem processor. Among them, the application processor mainly processes the operating system, user interface, application programs, etc., and the modem processor mainly processes wireless communication. It can be understood that the above modem processor may not be integrated into the processor 301.

[0207] The memory 302 can be used to store software programs and modules. The processor 301 executes various functional applications and data processing by running the software programs and modules stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area. Among them, the program storage area may store the operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area may store data created according to the use of the terminal. In addition, the memory 302 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices. Correspondingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.

[0208] The terminal also includes a power supply 303 that powers each component. In some embodiments, the power supply 303 may be logically connected to the processor 301 through a power management system, thereby implementing functions such as management of charging, discharging, and power consumption management through the power management system. The power supply 303 may also include any components such as one or more DC or AC power supplies, a recharge system, a power failure detection circuit, a power converter or inverter, a power status indicator, etc.

[0209] The terminal may further include an input module 304, which may be used to receive input numerical or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0210] The terminal may further include a communication module 305. In some embodiments, the communication module 305 may include a wireless module. The terminal may perform short-distance wireless transmission through the wireless module of the communication module 305, thereby providing users with wireless broadband Internet access. For example, the communication module 305 may be used to help users send and receive emails, browse web pages, access streaming media, etc.

[0211] Although not shown, the terminal may further include a display unit, etc., which will not be elaborated here. Specifically, in this embodiment, the processor 301 in the terminal will load the executable files corresponding to the processes of one or more application programs into the memory 302 according to the following instructions, and the processor 301 will run the application programs stored in the memory 302 to implement various functions as follows:

[0212] Retrieve at least two original data files from the data file database, and obtain the confidentiality level labels corresponding to the original data files, where the confidentiality level labels are used to represent one selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels;

[0213] Based on the confidentiality level labels, perform desensitization processing on the original data files to obtain processed original data files;

[0214] Through a preset confidentiality evaluation model, obtain the confidentiality evaluation result corresponding to the processed original data file according to the processed original data file;

[0215] Based on the confidentiality level labels and the confidentiality evaluation results, determine the desensitization deviation value corresponding to the original data file;

[0216] If there are desensitization deviation values not less than a first preset ratio among the desensitization deviation values that are less than a preset desensitization deviation threshold, determine that the processed original data file is a target data file.

[0217] [[ID=****]]The specific implementation of each of the above operations can be referred to the previous embodiments and will not be elaborated here.

[0218] As can be seen from the above, the embodiments of the present application can ensure that the data desensitization processing effect of the original data files in the same batch reaches the expected level, enabling files with different confidentiality levels to adapt to different data confidentiality effects, and continuously enhancing the data desensitization effect by quantifying the error of the desensitization processing, thereby improving the overall confidentiality management level of the data files.

[0219] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling relevant hardware through instructions. These instructions can be stored in a medium and loaded and executed by a processor.

[0220] For this reason, an embodiment of the present application provides a computer storage medium, in which multiple instructions are stored. These instructions can be loaded by a processor to execute the steps in any one of the data processing methods based on data desensitization provided by the embodiments of the present application. For example, the instructions can execute the following steps:

[0221] Retrieve at least two original data files from the data file database, and obtain the confidentiality level labels corresponding to the original data files. Among them, the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels;

[0222] Based on the confidentiality level label, perform desensitization processing on the original data file to obtain a processed original data file;

[0223] Through a preset confidentiality evaluation model, obtain a confidentiality evaluation result corresponding to the processed original data file according to the processed original data file;

[0224] Based on the confidentiality level label and the confidentiality evaluation result, determine the desensitization deviation value corresponding to the original data file;

[0225] If there are desensitization deviation values not less than a first preset ratio that are less than a preset desensitization deviation threshold, determine the processed original data file as the target data file.

[0226] Among them, the computer storage medium may include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disc, etc.

[0227] According to one aspect of the present application, a computer program product or a computer program is provided. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer storage medium. The processor of the computer device reads the computer instructions from the medium, and the processor executes the computer instructions, so that the computer device executes the methods provided in the various optional implementation manners provided in the above embodiments.

[0228] Since the instructions stored in the computer storage medium can execute the steps in any one of the data processing methods based on data desensitization provided by the embodiments of the present application, the beneficial effects achievable by any one of the data processing methods based on data desensitization provided by the embodiments of the present application can be realized. For details, refer to the previous embodiments and will not be elaborated herein.

[0229] The above has introduced in detail a data processing method, device, terminal, and medium based on data desensitization provided by the embodiments of the present application. Specific examples are used herein to elaborate on the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present application.

Claims

1. A data processing method based on data desensitization, characterized in that Applied to a data middle platform, the data middle platform includes a data file repository and a server, and both the data file repository and the server are in the same network; The method includes: Retrieve at least two original data files from the data file repository and obtain the confidentiality level labels corresponding to the original data files. Among them, the confidentiality level labels are used to represent one confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels; Based on the confidentiality level labels, perform desensitization processing on the original data files to obtain processed original data files; Through a preset confidentiality evaluation model, obtain the confidentiality evaluation result corresponding to the processed original data file according to the processed original data file; Based on the confidentiality level labels and the confidentiality evaluation results, determine the desensitization deviation value corresponding to the original data file; If there are desensitization deviation values not less than a first preset ratio that are less than the preset desensitization deviation threshold, determine the processed original data file as the target data file.

2. The method according to claim 1, wherein A preset desensitization processing model set is loaded in the server, and the preset desensitization processing model set includes at least two different preset desensitization processing models, and the preset desensitization processing models correspond to the confidentiality levels one by one; The step of performing desensitization processing on the original data file based on the confidentiality level label to obtain the processed original data file includes: Based on the confidentiality level label, select the preset desensitization processing model corresponding to the confidentiality level label from the preset desensitization processing model set; Through the preset desensitization processing model corresponding to the confidentiality level label, obtain the processed original data file according to the original data file.

3. The method according to claim 2, characterized in that, The preset desensitization processing model includes a first input layer, a first embedding layer, a first convolutional layer, a pooling layer, a first fully connected layer, and a first output layer; The step of obtaining the processed original data file according to the original data file through the preset desensitization processing model corresponding to the confidentiality level label includes: Through the first input layer, obtain a first sensitive information tensor according to the sensitive information in the original data file; Through the first embedding layer, obtain a first floating-point tensor according to the first sensitive information tensor; Through the first convolutional layer, obtain a second floating-point tensor according to the first floating-point tensor; Through the pooling layer, obtain a third floating-point tensor according to the second floating-point tensor; Through the first fully connected layer, obtain a fourth floating-point tensor according to the third floating-point tensor; Through the first output layer, obtain a first output tensor according to the fourth floating-point tensor, where the first output tensor is used to represent the desensitized sensitive information; Based on the original data file and the first output tensor, obtain the processed original data file.

4. The method according to claim 2, wherein The preset desensitization processing model includes a second input layer, a second embedding layer, a second convolutional layer, a recurrent neural network layer, an attention mechanism layer, a second fully connected layer, and a second output layer; Obtaining the processed original data file according to the original data file through the preset desensitization processing model corresponding to the confidentiality level label includes: Obtaining a second sensitive information tensor through the second input layer according to the sensitive information in the original data file; Obtaining a fifth floating-point tensor through the second embedding layer according to the second sensitive information tensor; Obtaining a sixth floating-point tensor through the second convolutional layer according to the fifth floating-point tensor; Obtaining a seventh floating-point tensor through the recurrent neural network layer according to the sixth floating-point tensor; Obtaining an eighth floating-point tensor through the attention mechanism layer according to the seventh floating-point tensor; Obtaining a ninth floating-point tensor through the second fully-connected layer according to the eighth floating-point tensor; Obtaining a second output tensor through the second output layer according to the ninth floating-point tensor, where the second output tensor is used to represent the desensitized sensitive information; Obtaining the processed original data file based on the original data file and the second output tensor.

5. The method according to claim 1, characterized in that The preset confidentiality evaluation model includes a third input layer, a third embedding layer, a self-attention mechanism layer, a multi-layer perceptron layer, and a third output layer; Obtaining the confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through the preset confidentiality evaluation model includes: Obtaining an input tensor through the third input layer according to the processed original data file; Obtaining a tenth floating-point tensor through the third embedding layer according to the input tensor; Obtaining an eleventh floating-point tensor through the self-attention mechanism layer according to the tenth floating-point tensor; Obtaining a twelfth floating-point tensor through the multi-layer perceptron layer according to the eleventh floating-point tensor; Obtaining a confidentiality evaluation score through the third output layer according to the twelfth floating-point tensor, and determining the confidentiality evaluation score as the confidentiality evaluation result.

6. The method according to claim 5, wherein Determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result includes: Converting the confidentiality level corresponding to the confidentiality level label into a confidentiality level value corresponding to the confidentiality level label through a preset linear mapping function; Calculating the absolute value of the difference between the confidentiality level value and the confidentiality evaluation score to obtain the desensitization deviation value.

7. The method according to claim 2, characterized in that, After determining the desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result, the method further includes: If there are desensitization deviation values not less than a second preset ratio greater than a preset desensitization deviation threshold among the desensitization deviation values, updating the preset desensitization processing model based on the desensitization deviation values to obtain an updated preset desensitization processing model, where the updated preset desensitization model is used to desensitize the original data file.

8. A data processing device based on data desensitization, characterized in that, Applied to a data middle platform, the data middle platform includes a data file database and a server, and both the data file database and the server are in the same network; The device includes: A first unit for retrieving at least two original data files from a data file repository and obtaining a confidentiality level label corresponding to the original data file, where the confidentiality level label is used to represent a confidentiality level selected from a preset confidentiality level set for the original data file, and the preset confidentiality level set includes at least two different confidentiality levels; A second unit for performing desensitization processing on the original data file based on the confidentiality level label to obtain a processed original data file; A third unit for obtaining a confidentiality evaluation result corresponding to the processed original data file according to the processed original data file through a preset confidentiality evaluation model; A fourth unit for determining a desensitization deviation value corresponding to the original data file based on the confidentiality level label and the confidentiality evaluation result; A fifth unit for determining the processed original data file as a target data file if there are desensitization deviation values not less than a first preset ratio that are less than a preset desensitization deviation threshold.

9. A terminal, characterized in that, It includes a processor and a memory, and the memory stores multiple instructions; the processor loads the instructions from the memory to execute the steps in the method according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The computer storage medium stores multiple instructions, and the instructions are suitable for being loaded by the processor to execute the steps in the method according to any one of claims 1 to 7.

Images