Joint factoring data security protection method and system based on block chain

Abstract

The embodiment of the invention provides a joint factoring data security protection method and system based on a block chain, and belongs to the technical field of data transmission. The method comprises the steps of performing key statement recognition on initial factoring data to obtain a target key statement; performing sensitive information identification on the target key statement to obtain target sensitive information; performing data entropy analysis according to the initial factoring data and the target sensitive information to obtain a first encryption mode and a second encryption mode; performing encryption according to the first encryption mode and the second encryption mode to obtain a target encryption result; performing intention recognition on the initial factoring data to obtain a target intention, and determining a target permission corresponding to the target encryption result according to the target intention; transmitting the target encryption result and the target permission to a related node by using a block chain, and obtaining a related permission corresponding to the related node; and screening the related nodes according to the target authority and the related authority to obtain a target node, so that the target object obtains a target encryption result according to the target node.

Description

Technical Field

[0001] The present invention relates to the technical field of data transmission, and particularly to a method and system for protecting the data security of joint factoring based on blockchain. Background Art

[0002] In today's digital age, the popularization of Internet-based electronic office work and the rapid development of Internet finance have made enterprises' demand for the electronic informatization of factoring services become increasingly urgent. Traditional joint factoring business mainly relies on offline operation models. In key links such as customer review, credit limit approval, and financing review, business personnel on both sides need to conduct frequent offline communication and review to verify the authenticity and validity of the materials and information provided by customers at each stage.

[0003] However, this offline business processing method has many limitations and challenges. First of all, due to the lack of a unified electronic platform to manage the business materials and information of both parties or multiple parties, it is difficult to ensure the consistency and accuracy of data. Secondly, offline communication not only consumes a large amount of human resources, but also increases communication costs and reduces the efficiency of business processing. In addition, this traditional operation model also increases the probability of potential risks, such as information leakage, review negligence and other issues.

[0004] Therefore, in order to adapt to the development of the times and meet the market demand, there is an urgent need for an electronic informatization solution for joint factoring business to improve the efficiency of business processing, reduce operating costs, and enhance the ability of risk management. Summary of the Invention

[0005] The main purpose of the embodiments of the present invention is to provide a method and system for protecting the data security of joint factoring based on blockchain, aiming to solve the problems of low efficiency and high risk in processing joint factoring business in related technologies.

[0006] In a first aspect, the embodiments of the present invention provide a method for protecting the data security of joint factoring based on blockchain, including:

[0007] Obtain the initial factoring data corresponding to the target construction enterprise, and perform key sentence recognition on the initial factoring data to obtain the corresponding target key sentences;

[0008] Perform sensitive information recognition on the target key sentences to obtain the target sensitive information corresponding to the target key sentences;

[0009] Perform data entropy analysis according to the initial factoring data and the target sensitive information to obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data;

[0010] Encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result;

[0011] Perform intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data, and determine the target permission corresponding to the target encryption result according to the target intent;

[0012] Transmit the target encryption result and the target permission to relevant nodes using a blockchain, and obtain the relevant permissions corresponding to the relevant nodes;

[0013] Screen the relevant nodes according to the target permission and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

[0014] In a second aspect, an embodiment of the present invention provides a blockchain-based joint factoring data security protection system, including:

[0015] A data acquisition module, configured to obtain the initial factoring data corresponding to a target construction enterprise, and perform key statement recognition on the initial factoring data to obtain the corresponding target key statement;

[0016] An information recognition module, configured to perform sensitive information recognition on the target key statement to obtain the target sensitive information corresponding to the target key statement;

[0017] A method determination module, configured to perform data entropy analysis according to the initial factoring data and the target sensitive information to obtain a first encryption method corresponding to the target sensitive information and a second encryption method corresponding to the initial factoring data;

[0018] A data encryption module, configured to encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result;

[0019] An intent recognition module, configured to perform intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data, and determine the target permission corresponding to the target encryption result according to the target intent;

[0020] A data transmission module, configured to transmit the target encryption result and the target permission to relevant nodes using a blockchain, and obtain the relevant permissions corresponding to the relevant nodes;

[0021] A data screening module, configured to screen the relevant nodes according to the target permission and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

[0022] An embodiment of the present invention provides a method and system for protecting the data security of joint factoring based on blockchain. The method includes: obtaining the initial factoring data corresponding to the target construction enterprise, and identifying the key sentences in the initial factoring data to obtain the corresponding target key sentences. Furthermore, by identifying the key sentences in the initial factoring data, the most important information can be extracted, reducing the processing of irrelevant data and lowering the risk of data leakage. Then, identify the sensitive information in the target key sentences to obtain the target sensitive information corresponding to the target key sentences, perform data entropy analysis based on the initial factoring data and the target sensitive information, obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data, and thus select an appropriate encryption method to encrypt the sensitive information and the initial factoring data according to the data entropy analysis, further ensuring the confidentiality and integrity of the data. Then, encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain the target encryption result. Furthermore, identify the intention of the initial factoring data to obtain the target intention corresponding to the initial factoring data, and determine the target permission corresponding to the target encryption result according to the target intention. Thus, using blockchain technology to transmit the encrypted data can ensure the immutability and transparency of data transmission, preventing the data from being maliciously tampered with or stolen during the transmission process. Furthermore, when the data is transmitted to the relevant nodes, further verify the permissions of the nodes to ensure that the data recipient has legitimate access permissions. Then, screen the relevant nodes according to the target permission and the relevant permissions to obtain the target nodes, so that the target object can obtain the target encryption result according to the target nodes, that is, the target object obtains the corresponding target encryption result. Thus, it can effectively reduce the risk of data leakage, improve the automation level of data processing, and further improve the efficiency of data transmission. It also solves the problems of low efficiency and high risk in processing joint factoring business in the related art. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0024] Figure 1 It is a schematic flowchart of a method for protecting the data security of joint factoring based on blockchain provided by an embodiment of the present invention;

[0025] Figure 2 It is a schematic module structure diagram of a system for protecting the data security of joint factoring based on blockchain provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0027] The flowchart shown in the accompanying drawings is only an example, and does not necessarily include all the contents and operations / steps, nor does it necessarily need to be executed in the described order. For example, some operations / steps can be decomposed, combined or partially merged, so the actual execution order may change according to the actual situation.

[0028] It should be understood that the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0029] The embodiments of the present invention provide a method and system for data security protection of joint factoring based on blockchain. Among them, the method for data security protection of joint factoring based on blockchain can be applied to a terminal device, and the terminal device can be an electronic device such as a tablet computer, a notebook computer, a desktop computer, a personal digital assistant, and a wearable device. The terminal device can be a server or a server cluster.

[0030] Next, some embodiments of the present invention will be described in detail in conjunction with the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0031] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a method for data security protection of joint factoring based on blockchain provided by an embodiment of the present invention.

[0032] As Figure 1 shown, the method for data security protection of joint factoring based on blockchain includes steps S101 to S107.

[0033] Step S101, obtain the initial factoring data corresponding to the target construction enterprise, and identify the key sentences of the initial factoring data to obtain the corresponding target key sentences.

[0034] Exemplarily, in the construction industry, factoring business is a comprehensive financial service that takes the transfer of the creditor's accounts receivable as a prerequisite and integrates financing, accounts receivable collection, management, and bad debt guarantee. To better adapt to the diversified development of the market economy, the commercial factoring model has become increasingly diverse. In addition to the traditional forward and reverse factoring models, the joint factoring model is also increasingly favored by construction enterprises. Its advantage lies in being able to provide large-scale financing, realizing the complementary advantages of assets and funds between different factoring companies, improving the credit system network while keeping warm in a group, achieving resource sharing, and broadening the financing channels of factoring companies.

[0035] Exemplarily, obtain the initial factoring data required by the target construction enterprise when commercial factoring is needed. The initial factoring data includes, but is not limited to, information related to the content evaluation of commercial factoring, such as the unified social credit code, legal person qualification, credit report, etc. corresponding to the target construction enterprise. The initial factoring data is the core foundation of commercial factoring business, and its integrity and accuracy directly affect the risk assessment and business decision-making of factoring companies. The initial factoring data is used to represent the enterprise data required for the factoring company to conduct risk assessment on the target construction enterprise.

[0036] Exemplarily, perform semantic analysis on the text in the initial factoring data to ensure the accuracy and effectiveness of the statements. For example, determine whether the statements involve core business data (such as accounts receivable, contract terms, etc.), and then assign weights to each selected key statement. The weights can be calculated according to the importance of the statements (such as whether they involve amounts, dates, contract terms, etc.), and finally select the target key statements with higher weights.

[0037] In some embodiments, the identifying the target key statements from the initial factoring data includes: performing word vector representation processing on the initial factoring data by using the word representation layer of the statement classification model to obtain initial word vectors; performing character vector representation processing on the initial factoring data by using the character-level representation layer of the statement classification model to obtain initial character vectors; performing aggregation processing on the initial word vectors by using the target convolutional layer of the statement classification model to obtain target convolutional results; performing high-level word-level feature extraction and information fusion processing on the target convolutional results by using the cross-channel fusion layer of the statement classification model to obtain initial semantic features; performing prominent feature screening on the initial semantic features by using the multi-head attention layer of the statement classification model to obtain target semantic features; performing fusion processing on the initial character vectors and the target semantic features by using the gated fusion layer of the statement classification model to obtain target fusion features; performing type classification on the target fusion features by using the statement classification layer of the statement classification model to obtain the target statement type corresponding to the initial factoring data; and identifying the target key statements from the initial factoring data according to the target statement type.

[0038] Exemplarily, the word representation layer of the statement classification model is used to perform word vector representation processing on the initial factoring data to obtain initial word vectors. By using a pre-trained word embedding model (such as Word2Vec, GloVe) or a pre-trained language model such as BERT, the words in the initial factoring data are converted into vector representations of a fixed dimension.

[0039] Exemplarily, the character-level representation layer of the statement classification model is used to perform character vector representation processing on the initial factoring data to obtain initial character vectors, that is, a pre-trained character embedding model is used to convert each character in the initial factoring data into a vector representation.

[0040] Exemplarily, the target convolution layer of the statement classification model is used to perform aggregation processing on the initial word vectors to obtain target convolution results. For example, a one-dimensional convolutional neural network (CNN) is used to slide on the word vector sequence to capture local features and generate higher-level feature representations.

[0041] Exemplarily, the cross-channel fusion layer of the statement classification model is used to perform fusion processing on the target convolution results for high-level word-level features to obtain initial semantic features. For example, semantic information is extracted from the convolution features in multiple ways (such as max pooling, average pooling) and fused.

[0042] Exemplarily, the multi-head attention layer of the statement classification model is used to screen out prominent features from the initial semantic features to obtain target semantic features. Furthermore, the attention mechanism is used to emphasize the feature parts that contribute more to the statement meaning and suppress the unimportant parts.

[0043] Exemplarily, the gated fusion layer of the statement classification model is used to perform fusion processing on the initial character vectors and the target semantic features to obtain target fusion features. Through the gating mechanism, the weights of the character vectors and the semantic features are dynamically adjusted to achieve effective feature fusion. Finally, the statement classification layer is used to perform type classification on the target fusion features to obtain the target statement type corresponding to the initial factoring data. For example, through a fully connected layer and a softmax activation function, the fusion features are mapped to predefined statement types for classification to obtain the corresponding target statement type. The target statement type is either a core statement or a non-core statement.

[0044] Exemplarily, the sentence with the target statement type being a core statement is determined as the target key sentence corresponding to the initial factoring data. Thus, based on the classification result, the sentences belonging to the key statement type are screened out. The target key sentences generally contain important information such as amount, date, contract terms, etc.

[0045] Step S102: Identify sensitive information from the target key sentence to obtain the target sensitive information corresponding to the target key sentence.

[0046] Exemplarily, the target key statement is cleaned and standardized. For example, irrelevant characters such as punctuation marks and special symbols in the target key statement are deleted, and the sentence is tokenized into words or phrases for subsequent processing. Define the categories of sensitive information to be recognized. Common categories of sensitive information include, but are not limited to, names, phone numbers, addresses, ID numbers, bank account numbers, contract amounts, dates, etc. Then, a rule library containing relevant regular expressions or keyword lists is established for each category. Thus, rule matching technology and natural language processing technology are used to identify sensitive information in the target key statement, and then the target sensitive information corresponding to the target key statement is obtained.

[0047] In some embodiments, the obtaining of the target sensitive information corresponding to the target key statement by identifying sensitive information in the target key statement includes: performing named entity recognition on the target key statement to obtain the first entity type corresponding to the target word in the target key statement; determining a preset type and preset words corresponding to the preset type, and replacing the target word in the target key statement with the preset words according to the preset type and the first entity type to obtain a target conversion statement corresponding to the target key statement; performing word frequency statistics on the target conversion statement to obtain the word weights corresponding to the entity type in the target conversion statement; determining a sensitive score corresponding to the target conversion statement according to the word weights, and screening out the target sensitive statement corresponding to the target key statement according to the sensitive score; obtaining the second entity type corresponding to the target sensitive statement from the first entity type, and determining the target sensitive information corresponding to the target sensitive statement according to the second entity type.

[0048] Exemplarily, named entity recognition is performed on the target key statement to obtain the first entity type corresponding to the target word. A pre-trained named entity recognition model is used to identify entities in the target key statement and label their types (such as person names, place names, dates, amounts, etc.), and then the identified entities are labeled as the first entity type.

[0049] Exemplarily, a preset type and its corresponding preset words are determined, and the target word in the target key statement is replaced according to the preset type and the first entity type to obtain a target conversion statement. For example, the entity types to be replaced are defined, and corresponding preset words are established for the preset type, such as replacing the preset type with a unified fixed identifier. Thus, according to the first entity type, the target word in the target key statement is replaced with the preset words. For example, the person name "Zhang San" is replaced with MN, and the time "October 1, 2023" is also replaced with MN.

[0050] Exemplarily, perform word frequency statistics on the target conversion statement to obtain the word weights corresponding to the entity types in the target conversion statement. For example, count the occurrence frequencies of each preset word in the target conversion statement, and then calculate the word weights for each of the entire preset types based on the word frequencies.

[0051] Exemplarily, determine the sensitivity score corresponding to the target conversion statement based on the word weights, and filter out the target sensitive statements corresponding to the target key statements according to the sensitivity score. For example, calculate the sensitivity score of the target conversion statement based on the word weights. For example, the sensitivity score can be obtained by weighted summation based on the weights of the entity types, and then set a threshold for the sensitivity score, and filter out the statements with sensitivity scores higher than the threshold as the target sensitive statements.

[0052] Exemplarily, obtain the second entity type corresponding to the target sensitive statement from the first entity type, and determine the target sensitive information corresponding to the target sensitive statement according to the second entity type. For example, extract the second entity type (such as "person's name", "date", etc.) from the target sensitive statement, and then determine the corresponding target sensitive information in the target sensitive statement according to the second entity type. For example, if the second entity type is "person's name", the target sensitive information is the specific person's name (such as "Zhang San").

[0053] In some embodiments, the determining the sensitivity score corresponding to the target conversion statement according to the word weights includes: obtaining the entity position distribution corresponding to the entity type, and determining the position weight corresponding to the entity type according to the entity position distribution; determining the entity sensitivity level corresponding to the entity type according to a preset rule, and determining the type sensitivity weight corresponding to the entity type according to the entity sensitivity level; fusing the word weight, the position weight, and the type sensitivity weight to determine the entity sensitivity weight corresponding to the entity type; fusing the entity sensitivity weights to determine the sensitivity score corresponding to the target conversion statement; wherein, the sensitivity score is obtained according to the following formula:

[0054]

[0055] wherein, represents the sensitivity score corresponding to the i-th target conversion statement, represents the quantity corresponding to the entity type of the i-th target conversion statement, represents the word weight corresponding to the j-th entity type, represents the word frequency information corresponding to the j-th entity type, represents the position weight corresponding to the j-th entity type, represents the type sensitivity weight corresponding to the j-th entity type, 、 and Represent the adjustment parameters corresponding to the word weight, the position weight, and the type sensitivity weight respectively.

[0056] Exemplarily, obtain the entity position distribution corresponding to each entity type, and determine the position weight corresponding to the entity type according to the entity position distribution. For example, count the position information of each entity type in the target conversion statement, such as whether the position where the entity appears is at the beginning of the sentence, in the middle of the sentence, or at the end of the sentence, and assign different weights according to the importance of the position. Generally, the entities at the beginning and end of the sentence may be more important and can be assigned higher weights. For example, the position weight at the beginning of the sentence is 1.2, in the middle of the sentence is 1.0, and at the end of the sentence is 1.1, so as to calculate the average position weight of each entity type in all target conversion statements as the position weight of the entity type.

[0057] Exemplarily, determine the entity sensitivity level corresponding to the entity type according to the preset rules, and determine the type sensitivity weight corresponding to the entity type according to the entity sensitivity level. For example, define the sensitivity levels of different entity types, such as "high", "medium", and "low". For example, when the entity type is amount, the sensitivity level is high; when the entity type is date, the sensitivity level is medium; when the entity type is location, the sensitivity level is low, and map the sensitivity level to specific weight values, such as "high" is 1.5, "medium" is 1.0, "low" is 0.5, so as to determine the type sensitivity weight of the entity type according to the weight value corresponding to the sensitivity level.

[0058] Exemplarily, fuse the word weight, the position weight, and the type sensitivity weight according to the following formula to determine the entity sensitivity weight corresponding to the entity type, and determine the sensitivity score corresponding to the target conversion statement according to the entity sensitivity weight:

[0059]

[0060] Among them, represents the sensitivity score corresponding to the i-th target conversion statement, represents the quantity corresponding to the entity type corresponding to the i-th target conversion statement, represents the word weight corresponding to the j-th entity type, represents the word frequency information corresponding to the j-th entity type, represents the position weight corresponding to the j-th entity type, represents the type sensitivity weight corresponding to the j-th entity type, , and represent the adjustment parameters corresponding to the word weight, the position weight, and the type sensitivity weight respectively.

[0061] Exemplarily, by integrating word weight, position weight, and type-sensitive weight, the word frequency, position distribution, and sensitivity level of entities in the statement are comprehensively considered, making the calculation of sensitivity scores more comprehensive and accurate. Compared with the analysis of a single dimension (such as word frequency), integrating multi-dimensional information can more accurately reflect the contribution of entities to the overall sensitivity of the statement.

[0062] Exemplarily, the above formula introduces adjustment parameters , and , which can flexibly adjust the importance of word weight, position weight, and type-sensitive weight according to actual needs. The use of adjustment parameters enables flexible weight allocation in different scenarios to adapt to different business requirements. For example, in some scenarios, more emphasis can be placed on word frequency, then is increased. In other scenarios, more emphasis is placed on position or sensitivity level, then and are increased.

[0063] Exemplarily, through the weighted sum formula, the sensitivity weight of each entity type can be accurately calculated, and then the sensitivity score of the target conversion statement can be accurately calculated. The improvement in the accuracy of the sensitivity score helps to more precisely screen out highly sensitive statements, avoid misjudgment or missed judgment, and makes the calculation of the sensitivity score more scientific and efficient, providing strong support for subsequent sensitive information screening and management.

[0064] Step S103: Perform data entropy analysis based on the initial factoring data and the target sensitive information to obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data.

[0065] Exemplarily, existing publicly available data is obtained, and data entropy analysis is performed on the initial factoring data and the target sensitive information under the existing publicly available data, so as to obtain the randomness corresponding to the initial factoring data and the target sensitive information respectively.

[0066] Exemplarily, if the data entropy of the target sensitive information is high, indicating that the target sensitive information is relatively random, a symmetric encryption algorithm (such as AES) can be considered because symmetric encryption is efficient and suitable for processing data with high randomness. If the data entropy is low, indicating that the data has a certain regularity, a more secure encryption algorithm, such as asymmetric encryption (such as RSA), or a hybrid encryption method combining symmetric and asymmetric encryption can be considered. Similarly, for the initial factoring data with high entropy, an efficient symmetric encryption algorithm is selected, and for low-entropy data, more advanced encryption means may be required to protect the security of the data, such as using a longer key or multiple encryption techniques.

[0067] In some embodiments, performing data entropy analysis based on the initial factoring data and the target sensitive information to obtain a first encryption method corresponding to the target sensitive information and a second encryption method corresponding to the initial factoring data includes: removing the target sensitive information from the initial factoring data to obtain target factoring data; calculating the information entropy of the target sensitive information to obtain a first entropy value corresponding to the target sensitive information; calculating the information entropy of the target factoring data to obtain a second entropy value corresponding to the target factoring data; determining a first preset entropy value corresponding to the target sensitive information and a second preset entropy value corresponding to the target factoring data; determining a first information frequency level corresponding to the target sensitive information according to the first entropy value and the first preset entropy value; determining the first encryption method corresponding to the target sensitive information according to the first information frequency level; determining a second information frequency level corresponding to the target factoring data according to the second entropy value and the second preset entropy value; and determining the second encryption method corresponding to the target factoring data according to the second information frequency level.

[0068] Exemplarily, the target sensitive information is removed from the initial factoring data to obtain target factoring data. For example, the target sensitive information is identified and located by keyword matching, regular expressions or other methods, so that the identified target sensitive information is removed from the initial factoring data to obtain target factoring data.

[0069] Exemplarily, the information entropy of the target sensitive information and the target factoring data is calculated to obtain a first entropy value and a second entropy value respectively, and then the threshold values of the entropy values are set according to historical data or industry standards to determine the first preset entropy value of the target sensitive information and the second preset entropy value of the target factoring data.

[0070] Exemplarily, the first information frequency level of the target sensitive information is determined by comparing the first entropy value and the first preset entropy value. And the second information frequency level of the target factoring data is determined by comparing the second entropy value and the second preset entropy value. For example, if the first entropy value is higher than the first preset entropy value, the first information frequency level is low; otherwise it is high.

[0071] Exemplarily, according to the first information frequency level, the first encryption method corresponding to the target sensitive information is determined, and the first encryption method can be either dynamic encryption or static encryption. At the same time, according to the second information frequency level, the second encryption method corresponding to the target factoring data is determined, and the second encryption method can also be either dynamic encryption or static encryption. Among them, static encryption means encrypting data with a fixed encryption password, while dynamic encryption means encrypting with different encryption passwords to enhance security. In this way, the most suitable encryption method can be selected according to the sensitivity and frequency level of the data to ensure the security of information and the efficiency of the system.

[0072] Step S104: Encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result.

[0073] Exemplarily, encrypt the target sensitive information using the first encryption method to obtain a first encryption result, and encrypt the initial factoring data using the second encryption method to obtain a second encryption result, so as to merge the first encryption result and the second encryption result to obtain the target encryption result.

[0074] In some embodiments, the step of encrypting the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result includes: when the first encryption method is dynamic encryption, determine the first classification attribute of the target sensitive information in the first encryption key, and determine the first code corresponding to the target sensitive information according to the first classification attribute; determine the first key information corresponding to the target sensitive information in the first encryption key according to the first code and the first entropy value; when the first encryption method is static encryption, obtain the first key information corresponding to the target sensitive information from the database; encrypt the target sensitive information according to the first key information to obtain a first encryption result; when the second encryption method is dynamic encryption, determine the second classification attribute of the target factoring data in the second encryption key, and determine the second code corresponding to the target factoring data according to the second classification attribute; determine the second key information corresponding to the target factoring data in the second encryption key according to the second code and the second entropy value; when the second encryption method is static encryption, obtain the second key information corresponding to the target factoring data from the database; encrypt the target factoring data according to the second key information to obtain a second encryption result; fuse the first encryption result and the second encryption result to obtain the target encryption result.

[0075] Exemplarily, determine whether the first encryption method is dynamic encryption or static encryption. When the first encryption method is dynamic encryption, determine the classification attribute of the target sensitive information in the first encryption key. For example, when the target sensitive information is a number, the first classification attribute is the number type, and when the target sensitive information is text, the first classification attribute is the text type. Thus, according to different first classification attributes, use the corresponding coding type to obtain the first code corresponding to the target sensitive information, and then combine the first code and the first entropy value to determine the first key information corresponding to the target sensitive information in the first encryption key, so as to encrypt the target sensitive information according to the first key information to obtain the first encryption result.

[0076] Exemplarily, when the first encryption method is static encryption, the first key information corresponding to the target sensitive information is directly obtained from the database, and then the target sensitive information is encrypted using the determined first key information to obtain a first encryption result.

[0077] Exemplarily, it is determined whether the second encryption method is dynamic encryption or static encryption. When the second encryption method is dynamic encryption, the classification attribute of the target factoring information in the second encryption key is determined. For example, when the target factoring information is a number, the second classification attribute is the numeric type, and when the target factoring information is text, the second classification attribute is the text type. Then, according to different second classification attributes, the second encoding corresponding to the target factoring information is obtained using the corresponding encoding type. Furthermore, the second key information corresponding to the target sensitive information in the second encryption key is determined by combining the second encoding and the second entropy value. Thus, the target sensitive information is encrypted according to the second key information to obtain a second encryption result.

[0078] Exemplarily, when the second encryption method is static encryption, the second key information corresponding to the target factoring information is directly obtained from the database, and then the target factoring information is encrypted using the determined second key information to obtain a second encryption result.

[0079] Exemplarily, the first encryption result and the second encryption result are fused using a splicing algorithm to generate a final target encryption result.

[0080] In some embodiments, determining the first key information corresponding to the target sensitive information in the first encryption key according to the first encoding and the first entropy value includes: determining a first maximum entropy value corresponding to the target sensitive information from the first entropy value; determining a first key distribution parameter corresponding to the target sensitive information according to the first entropy value and the first maximum entropy value; determining a first frequency corresponding to the target sensitive information, and adjusting the first encoding according to the first frequency and the first key distribution parameter to obtain a first target encoding; and classifying the first target encoding using a first classification model to obtain the first key information corresponding to the target sensitive information in the first encryption key.

[0081] Exemplarily, by analyzing the first entropy value, the maximum value therein is found, and then this maximum value is determined as the first maximum entropy value for subsequent calculations. Then, according to the first entropy value and the first maximum entropy value, the first key distribution parameter corresponding to the target sensitive information is determined. For example, the ratio of the first entropy value to the first maximum entropy value is calculated to reflect the distribution characteristics of the data, and then based on the calculated ratio, the first key distribution parameter is determined. These parameters will be used to describe the distribution characteristics of the key.

[0082] Exemplarily, the occurrence frequencies of the elements in the target sensitive information are counted, and these frequencies are determined for subsequent coding adjustment. Thus, according to the first frequency and the first key distribution parameter, the first coding is adjusted to obtain the first target coding. For example, according to the first frequency and the first key distribution parameter, the structure and content of the first coding are adjusted to generate the adjusted first target coding to adapt to subsequent key classification.

[0083] Exemplarily, the first target coding is classified by the first classification model to obtain the first key information corresponding to the target sensitive information in the first encryption key.

[0084] Exemplarily, by analyzing the entropy value and frequency, the key distribution parameter and coding adjustment are accurately determined, improving the accuracy of key classification. Thus, the coding is adjusted according to different data characteristics, enhancing flexibility and adaptability. Furthermore, the first classification model is used for key classification, improving the acquisition efficiency of key information.

[0085] In some embodiments, the determining the second key information corresponding to the target factoring data in the second encryption key according to the second coding and the second entropy value includes: determining the second maximum entropy value corresponding to the target factoring data from the second entropy value; determining the second key distribution parameter corresponding to the target factoring data according to the second entropy value and the second maximum entropy value; determining the second frequency corresponding to the target factoring data, and adjusting the second coding according to the second frequency and the second key distribution parameter to obtain the second target coding; classifying the second target coding by the second classification model to obtain the second key information corresponding to the target factoring data in the second encryption key.

[0086] Exemplarily, by analyzing the second entropy value, the maximum value therein is found, and then this maximum value is determined as the second maximum entropy value for subsequent calculation. Then, according to the second entropy value and the second maximum entropy value, the second key distribution parameter corresponding to the target factoring information is determined. For example, the ratio of the second entropy value to the second maximum entropy value is calculated to reflect the distribution characteristics of the data, and then based on the calculated ratio, the second key distribution parameter is determined, and these parameters will be used to describe the distribution characteristics of the key.

[0087] Exemplarily, the occurrence frequencies of the elements in the target factoring information are counted, and these frequencies are determined for subsequent coding adjustment. Thus, according to the second frequency and the second key distribution parameter, the second coding is adjusted to obtain the second target coding. For example, according to the second frequency and the second key distribution parameter, the structure and content of the second coding are adjusted to generate the adjusted second target coding to adapt to subsequent key classification.

[0088] Exemplarily, the second classification model is used to classify the second target encoding to obtain the first key information corresponding to the target sensitive information in the second encryption key.

[0089] Exemplarily, by analyzing the entropy value and frequency, the key distribution parameters and encoding adjustments are accurately determined, improving the accuracy of key classification. Thus, the encoding is adjusted according to different data characteristics, enhancing flexibility and adaptability. Furthermore, the second classification model is used for key classification, improving the acquisition efficiency of key information.

[0090] Step S105: Perform intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data, and determine the target permission corresponding to the target encryption result according to the target intent.

[0091] Exemplarily, an intent recognition model (such as a natural language processing model or a machine learning model) is applied to perform intent recognition on the initial factoring data to determine the target intent of the initial factoring data, and the target intent is used to determine the receiving object corresponding to the initial factoring data.

[0092] Exemplarily, a mapping relationship between the target intent and the permission is established to clarify the permission levels corresponding to different intents. Thus, according to the identified target intent, the corresponding permission level is searched to determine the target permission of the target encryption result.

[0093] In some embodiments, the performing intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data includes: using the target recognition layer of the target intent model to perform key target recognition on the initial factoring data to obtain the target key information corresponding to the initial factoring data; using the text representation layer of the target intent model to perform data representation on the target key information to obtain an initial representation vector; using the relationship recognition layer of the target intent model to perform relationship recognition on the target key information according to the initial factoring data to obtain target relationship information; using the feature fusion layer of the target intent model to perform feature fusion on the initial representation vector according to the target relationship information to obtain a target representation vector; using the intent classification layer of the target intent model to perform intent recognition on the target representation vector to obtain the target intent corresponding to the initial factoring data.

[0094] Exemplarily, the target recognition layer of the target intent model is applied to extract the target key information from the initial factoring data. The target recognition layer typically uses deep learning techniques (such as convolutional neural networks or recurrent neural networks) to extract important entities and keywords in the text, and then uses the text representation layer of the target intent model to perform data representation on the target key information to obtain the initial representation vector. Features are extracted from the target key information, and these features can be word embeddings (such as Word2Vec, GloVe) or sentence embeddings (such as BERT), and then the extracted features are converted into vector form to generate the initial representation vector. This step typically uses word embedding techniques and context encoders (such as Transformer).

[0095] Exemplarily, the relationship recognition layer is applied to analyze the relationships between the target key information. The relationship recognition layer typically uses graph neural networks (GNNs) or attention mechanisms to identify the relationships between entities to obtain the target relationship information, and thus represents the identified target relationship information as structured data, such as a graph structure or a relationship matrix.

[0096] Exemplarily, the feature fusion layer of the target intent model performs feature fusion on the initial representation vector according to the target relationship information by means of concatenation, weighted average, or self-attention mechanism, etc., to obtain the target representation vector. Thus, an intent classification layer such as logistic regression, support vector machine, or deep neural network is used to classify the target representation vector, and then according to the classification result, the target intent corresponding to the initial factoring data is determined.

[0097] Step S106: Transmit the target encryption result and the target permission to relevant nodes using a blockchain, and obtain the relevant permissions corresponding to the relevant nodes.

[0098] Exemplarily, the target encryption result and the corresponding target permission are packaged into a data unit to ensure the integrity of the data, and then the packaged data is encoded, usually in JSON or Protobuf format, for easy storage and transmission on the blockchain. Thus, the target encryption result and the target permission are transmitted to relevant nodes using blockchain technology.

[0099] For example, select a suitable blockchain network (such as Ethereum, Hyperledger Fabric, etc.) to ensure the security and reliability of the network. Create a transaction on the blockchain that contains the target encryption result and the target permission. The transaction should also include information about the sender and the receiver, as well as the signature of the transaction (using the private key of the sender). Broadcast the created transaction to all nodes in the blockchain network for verification and recording.

[0100] Exemplarily, the relevant nodes query the recorded transactions through the blockchain network and extract the relevant permissions corresponding to the relevant nodes.

[0101] Step S107: Screen the relevant nodes according to the target permissions and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

[0102] Exemplarily, compare the target permissions with the relevant permissions and screen out the target nodes from the relevant nodes according to the comparison result, so that the target object can obtain the target encryption result according to the target nodes.

[0103] For example, analyze the specific contents of the target permissions and the relevant permissions, understand their permission scopes and levels, and then judge whether the relevant permissions of the relevant nodes meet the requirements of the target permissions. This can be carried out based on the inclusion relationship, priority, etc. of the permissions, and then determine the screening conditions. For example, the relevant permissions must completely include the target permissions, or the level of the relevant permissions is higher than or equal to the target permissions. Thus, according to the screening conditions, the nodes that meet the conditions are screened out from the relevant nodes as the target nodes. This ensures that the target object can only obtain the target encryption result from the target nodes and its access complies with the permission management regulations.

[0104] Please refer to Figure 2 , Figure 2A blockchain-based joint factoring data security protection system 200 provided by an embodiment of the present application. The blockchain-based joint factoring data security protection system 200 includes a data acquisition module 201, an information recognition module 202, a method determination module 203, a data encryption module 204, an intention recognition module 205, a data transmission module 206, and a data screening module 207. Among them, the data acquisition module 201 is used to obtain the initial factoring data corresponding to the target construction enterprise, and perform key statement recognition on the initial factoring data to obtain the corresponding target key statement; the information recognition module 202 is used to perform sensitive information recognition on the target key statement to obtain the target sensitive information corresponding to the target key statement; the method determination module 203 is used to perform data entropy analysis based on the initial factoring data and the target sensitive information to obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data; the data encryption module 204 is used to encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result; the intention recognition module 205 is used to perform intention recognition on the initial factoring data to obtain the target intention corresponding to the initial factoring data, and determine the target permission corresponding to the target encryption result according to the target intention; the data transmission module 206 is used to transmit the target encryption result and the target permission to relevant nodes through the blockchain, and obtain the relevant permissions corresponding to the relevant nodes; the data screening module 207 is used to screen the relevant nodes according to the target permission and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

[0105] In some embodiments, during the process of performing key statement recognition on the initial factoring data by the data acquisition module 201 to obtain the corresponding target key statement, the following operations are performed:

[0106] Perform word vector representation processing on the initial factoring data using the word representation layer of the sentence classification model to obtain initial word vectors;

[0107] Perform character vector representation processing on the initial factoring data using the character-level representation layer of the sentence classification model to obtain initial character vectors;

[0108] Perform aggregation processing on the initial word vectors using the target convolutional layer of the sentence classification model to obtain a target convolutional result;

[0109] Perform high-level word-level feature extraction information fusion processing on the target convolutional result using the cross-channel fusion layer of the sentence classification model to obtain initial semantic features;

[0110] Use the multi-head attention layer of the statement classification model to screen out prominent features from the initial semantic features to obtain target semantic features;

[0111] Use the gated fusion layer of the statement classification model to fuse the initial word vectors and the target semantic features to obtain target fusion features;

[0112] Use the statement classification layer of the statement classification model to classify the types of the target fusion features to obtain the target statement type corresponding to the initial factoring data;

[0113] Identify key statements from the initial factoring data according to the target statement type to obtain the target key statements.

[0114] In some embodiments, when the information recognition module 202 performs sensitive information recognition on the target key statements to obtain the target sensitive information corresponding to the target key statements, it executes:

[0115] Perform named entity recognition on the target key statements to obtain the first entity type corresponding to the target words in the target key statements;

[0116] Determine a preset type and preset words corresponding to the preset type, and replace the target words in the target key statements with the preset words according to the preset type and the first entity type to obtain a target conversion statement corresponding to the target key statements;

[0117] Perform word frequency statistics on the target conversion statements to obtain the word weights corresponding to the entity types in the target conversion statements;

[0118] Determine the sensitive score corresponding to the target conversion statement according to the word weights, and screen out the target sensitive statements corresponding to the target key statements according to the sensitive score;

[0119] Obtain the second entity type corresponding to the target sensitive statements from the first entity types, and determine the target sensitive information corresponding to the target sensitive statements according to the second entity type.

[0120] In some embodiments, when the information recognition module 202 determines the sensitive score corresponding to the target conversion statement according to the word weights, it executes:

[0121] Obtain the entity position distribution corresponding to the entity type, and determine the position weight corresponding to the entity type according to the entity position distribution;

[0122] Determine the entity sensitive level corresponding to the entity type according to preset rules, and determine the type sensitive weight corresponding to the entity type according to the entity sensitive level;

[0123] Fuse the word weight, the position weight, and the type sensitivity weight to determine the entity sensitivity weight corresponding to the entity type;

[0124] Fuse the entity sensitivity weight to determine the sensitivity score corresponding to the target conversion statement;

[0125] Among them, the sensitivity score is obtained according to the following formula:

[0126]

[0127] Among them, represents the sensitivity score corresponding to the i-th target conversion statement, represents the quantity corresponding to the entity type corresponding to the i-th target conversion statement, represents the word weight corresponding to the j-th entity type, represents the word frequency information corresponding to the j-th entity type, represents the position weight corresponding to the j-th entity type, represents the type sensitivity weight corresponding to the j-th entity type, , and represent the adjustment parameters corresponding to the word weight, the position weight, and the type sensitivity weight respectively.

[0128] In some embodiments, during the process of the method determination module 203 performing data entropy analysis on the initial factoring data and the target sensitive information to obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data, it performs:

[0129] Remove the target sensitive information from the initial factoring data to obtain target factoring data;

[0130] Perform information entropy calculation on the target sensitive information to obtain the first entropy value corresponding to the target sensitive information;

[0131] Perform information entropy calculation on the target factoring data to obtain the second entropy value corresponding to the target factoring data;

[0132] Determine the first preset entropy value corresponding to the target sensitive information and the second preset entropy value corresponding to the target factoring data;

[0133] Determine the first information frequency level corresponding to the target sensitive information according to the first entropy value and the first preset entropy value;

[0134] Determine the first encryption method corresponding to the target sensitive information according to the first information frequency level;

[0135] Determine the second information frequency level corresponding to the target factoring data according to the second entropy value and the second preset entropy value;

[0136] Determine the second encryption method corresponding to the target factoring data according to the second information frequency level.

[0137] In some embodiments, during the process of encrypting the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result, the data encryption module 204 performs:

[0138] When the first encryption method is dynamic encryption, determine the first classification attribute of the target sensitive information in the first encryption key, and determine the first encoding corresponding to the target sensitive information according to the first classification attribute;

[0139] Determine the first key information corresponding to the target sensitive information in the first encryption key according to the first encoding and the first entropy value;

[0140] When the first encryption method is static encryption, obtain the first key information corresponding to the target sensitive information from the database;

[0141] Encrypt the target sensitive information according to the first key information to obtain a first encryption result;

[0142] When the second encryption method is the dynamic encryption, determine the second classification attribute of the target factoring data in the second encryption key, and determine the second encoding corresponding to the target factoring data according to the second classification attribute;

[0143] Determine the second key information corresponding to the target factoring data in the second encryption key according to the second encoding and the second entropy value;

[0144] When the second encryption method is the static encryption, obtain the second key information corresponding to the target factoring data from the database;

[0145] Encrypt the target factoring data according to the second key information to obtain a second encryption result;

[0146] Fuse the first encryption result and the second encryption result to obtain the target encryption result.

[0147] In some embodiments, during the process of determining, according to the first encoding and the first entropy value, the first key information corresponding to the target sensitive information in the first encryption key, the data encryption module 204 performs the following:

[0148] Determine a first maximum entropy value corresponding to the target sensitive information from the first entropy value;

[0149] Determine a first key distribution parameter corresponding to the target sensitive information according to the first entropy value and the first maximum entropy value;

[0150] Determine a first frequency corresponding to the target sensitive information, and adjust the first encoding according to the first frequency and the first key distribution parameter to obtain a first target encoding;

[0151] Use a first classification model to perform key classification on the first target encoding to obtain the first key information corresponding to the target sensitive information in the first encryption key.

[0152] In some embodiments, during the process of determining, according to the second encoding and the second entropy value, the second key information corresponding to the target factoring data in the second encryption key, the data encryption module 204 performs the following:

[0153] Determine a second maximum entropy value corresponding to the target factoring data from the second entropy value;

[0154] Determine a second key distribution parameter corresponding to the target factoring data according to the second entropy value and the second maximum entropy value;

[0155] Determine a second frequency corresponding to the target factoring data, and adjust the second encoding according to the second frequency and the second key distribution parameter to obtain a second target encoding;

[0156] Use a second classification model to perform key classification on the second target encoding to obtain the second key information corresponding to the target factoring data in the second encryption key.

[0157] In some embodiments, during the process of performing intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data, the intent recognition module 205 performs the following:

[0158] Use the target recognition layer of the target intent model to perform key target recognition on the initial factoring data to obtain target key information corresponding to the initial factoring data;

[0159] Use the text representation layer of the target intent model to perform data representation on the target key information to obtain an initial representation vector;

[0160] The relationship recognition layer of the target intent model performs relationship recognition on the target key information according to the initial factoring data to obtain target relationship information;

[0161] The feature fusion layer of the target intent model performs feature fusion on the initial representation vector according to the target relationship information to obtain a target representation vector;

[0162] The intent classification layer of the target intent model performs intent recognition on the target representation vector to obtain the target intent corresponding to the initial factoring data.

[0163] In some embodiments, the blockchain-based joint factoring data security protection system 200 can be applied to a terminal device.

[0164] It should be noted that those skilled in the art can clearly understand that, for the convenience and conciseness of description, the specific working process of the above-described blockchain-based joint factoring data security protection system 200 can refer to the corresponding process in the foregoing embodiment of the blockchain-based joint factoring data security protection method, and will not be described herein again.

[0165] The embodiment of the present invention further provides a storage medium for computer-readable storage. The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any one of the blockchain-based joint factoring data security protection methods provided in the specification of the embodiment of the present invention.

[0166] Among them, the storage medium can be an internal storage unit of the terminal device in the foregoing embodiment, such as the hard disk or memory of the terminal device. The storage medium can also be an external storage device of the terminal device, such as a plug-in hard disk equipped on the terminal device, a smart media card (SMC), a secure digital (SD) card, a flash card, etc.

[0167] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations. In the hardware embodiments, the division between the functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, one physical component can have multiple functions, or one function or step can be executed by several physical components in cooperation. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cartridges, tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that a communication medium typically contains computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and can include any information delivery medium.

[0168] It should be understood that the term "and / or" used in the specification and claims of the present invention refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. It should be noted that in this document, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or system comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, article or system. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or system comprising the element.

[0169] The serial numbers of the embodiments of the present invention described above are only for description and do not represent the advantages or disadvantages of the embodiments. The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for data security protection of joint factoring based on blockchain, characterized in that, The method includes: Obtaining the initial factoring data corresponding to the target construction enterprise, and performing key sentence recognition on the initial factoring data to obtain the corresponding target key sentences; Performing sensitive information recognition on the target key sentences to obtain the target sensitive information corresponding to the target key sentences; Performing data entropy analysis based on the initial factoring data and the target sensitive information to obtain the first encryption method corresponding to the target sensitive information and the second encryption method corresponding to the initial factoring data; Encrypting the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result; Performing intention recognition on the initial factoring data to obtain the target intention corresponding to the initial factoring data, and determining the target permission corresponding to the target encryption result according to the target intention; Transmitting the target encryption result and the target permission to relevant nodes using a blockchain, and obtaining the relevant permissions corresponding to the relevant nodes; Screening the relevant nodes according to the target permission and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

2. The method according to claim 1, wherein The performing key sentence recognition on the initial factoring data to obtain the corresponding target key sentences includes: Performing word vector representation processing on the initial factoring data using the word representation layer of the sentence classification model to obtain an initial word vector; Performing character vector representation processing on the initial factoring data using the character-level representation layer of the sentence classification model to obtain an initial character vector; Performing aggregation processing on the initial word vector using the target convolutional layer of the sentence classification model to obtain a target convolutional result; Performing high-level word-level feature extraction information fusion processing on the target convolutional result using the cross-channel fusion layer of the sentence classification model to obtain an initial semantic feature; Performing prominent feature screening on the initial semantic feature using the multi-head attention layer of the sentence classification model to obtain a target semantic feature; Performing fusion processing on the initial character vector and the target semantic feature using the gated fusion layer of the sentence classification model to obtain a target fusion feature; Performing type classification on the target fusion feature using the sentence classification layer of the sentence classification model to obtain the target sentence type corresponding to the initial factoring data; Performing key sentence recognition on the initial factoring data according to the target sentence type to obtain the target key sentences.

3. The method according to claim 1, wherein The performing sensitive information recognition on the target key sentences to obtain the target sensitive information corresponding to the target key sentences includes: Performing named entity recognition on the target key sentences to obtain the first entity type corresponding to the target words in the target key sentences; Determining a preset type and the preset words corresponding to the preset type, and replacing the target words in the target key sentences according to the preset type and the first entity type with the preset words to obtain a target conversion sentence corresponding to the target key sentences; Performing word frequency statistics on the target conversion sentence to obtain the word weights corresponding to the entity type in the target conversion sentence; Determine the sensitivity score corresponding to the target conversion sentence according to the word weight, and filter out the target sensitive sentence corresponding to the target key sentence according to the sensitivity score; A second entity type corresponding to the target sensitive sentence is obtained from the first entity type, and the target sensitive information corresponding to the target sensitive sentence is determined according to the second entity type.

4. The method according to claim 3, wherein The determining the sensitivity score corresponding to the target conversion sentence according to the word weight includes: Obtaining entity position distribution corresponding to the entity type, and determining a position weight corresponding to the entity type according to the entity position distribution; Determine the entity sensitivity level corresponding to the entity type according to a preset rule, and determine the type sensitivity weight corresponding to the entity type according to the entity sensitivity level; The entity sensitive weight corresponding to the entity type is determined by fusing the word weight, the position weight and the type sensitive weight; The entity sensitivity weight is integrated to determine the sensitivity score corresponding to the target conversion statement; The sensitivity score is obtained according to the following formula: Among them, represents the sensitivity score corresponding to the i-th target conversion statement, represents the quantity corresponding to the entity type corresponding to the i-th target conversion statement, represents the word weight corresponding to the j-th entity type, represents the word frequency information corresponding to the j-th entity type, represents the position weight corresponding to the j-th entity type, represents the type sensitivity weight corresponding to the j-th entity type, , and represent the adjustment parameters corresponding to the word weight, the position weight, and the type sensitivity weight respectively.

5. The method according to claim 1, wherein The performing data entropy analysis according to the initial factoring data and the target sensitive information to obtain a first encryption method corresponding to the target sensitive information and a second encryption method corresponding to the initial factoring data includes: Eliminating the target sensitive information from the initial factoring data to obtain target factoring data; Performing information entropy calculation on the target sensitive information to obtain a first entropy value corresponding to the target sensitive information; Performing information entropy calculation on the target factoring data to obtain a second entropy value corresponding to the target factoring data; Determine a first preset entropy value corresponding to the target sensitive information and a second preset entropy value corresponding to the target factoring data; Determine a first information frequency level corresponding to the target sensitive information according to the first entropy value and the first preset entropy value; Determining a first encryption method corresponding to the target sensitive information according to the frequency level of the first information; Determining a second information frequency level corresponding to the target factoring data according to the second entropy value and the second preset entropy value; A second encryption method corresponding to the target factoring data is determined according to the second information frequency level.

6. The method according to claim 5, wherein The step of encrypting the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result includes: When the first encryption mode is dynamic encryption, determining a first classification attribute of the target sensitive information in the first encryption key, and determining a first code corresponding to the target sensitive information according to the first classification attribute; Determine first key information corresponding to the target sensitive information in the first encryption key according to the first code and the first entropy value; When the first encryption mode is static encryption, the first key information corresponding to the target sensitive information is obtained from a database; Encrypting the target sensitive information according to the first key information to obtain a first encryption result; When the second encryption method is the dynamic encryption, determine the second classification attribute of the target factoring data in the second encryption key, and determine the second encoding corresponding to the target factoring data according to the second classification attribute; Determine the second key information corresponding to the target factoring data in the second encryption key according to the second encoding and the second entropy value; When the second encryption method is the static encryption, obtain the second key information corresponding to the target factoring data from the database; Encrypt the target factoring data according to the second key information to obtain a second encryption result; Fuse the first encryption result and the second encryption result to obtain the target encryption result.

7. The method according to claim 6, wherein The determining the first key information corresponding to the target sensitive information in the first encryption key according to the first encoding and the first entropy value includes: Determine the first maximum entropy value corresponding to the target sensitive information from the first entropy value; Determine the first key distribution parameter corresponding to the target sensitive information according to the first entropy value and the first maximum entropy value; Determine the first frequency corresponding to the target sensitive information, and adjust the first encoding according to the first frequency and the first key distribution parameter to obtain a first target encoding; Use a first classification model to perform key classification on the first target encoding to obtain the first key information corresponding to the target sensitive information in the first encryption key.

8. The method according to claim 6, characterized in that, The determining the second key information corresponding to the target factoring data in the second encryption key according to the second encoding and the second entropy value includes: Determine the second maximum entropy value corresponding to the target factoring data from the second entropy value; Determine the second key distribution parameter corresponding to the target factoring data according to the second entropy value and the second maximum entropy value; Determine the second frequency corresponding to the target factoring data, and adjust the second encoding according to the second frequency and the second key distribution parameter to obtain a second target encoding; Use a second classification model to perform key classification on the second target encoding to obtain the second key information corresponding to the target factoring data in the second encryption key.

9. The method according to claim 1, characterized in that, The performing intent recognition on the initial factoring data to obtain the target intent corresponding to the initial factoring data includes: Use the target recognition layer of the target intent model to perform key target recognition on the initial factoring data to obtain the target key information corresponding to the initial factoring data; Use the text representation layer of the target intent model to perform data representation on the target key information to obtain an initial representation vector; Use the relationship recognition layer of the target intent model to perform relationship recognition on the target key information according to the initial factoring data to obtain target relationship information; Use the feature fusion layer of the target intent model to perform feature fusion on the initial representation vector according to the target relationship information to obtain a target representation vector; Use the intent classification layer of the target intent model to perform intent recognition on the target representation vector to obtain the target intent corresponding to the initial factoring data.

10. A blockchain-based joint factoring data security protection system, characterized in that, Includes: A data acquisition module, which is used to obtain the initial factoring data corresponding to the target construction enterprise, and identify key sentences in the initial factoring data to obtain corresponding target key sentences; An information recognition module, which is used to recognize sensitive information in the target key sentences to obtain target sensitive information corresponding to the target key sentences; A method determination module, which is used to perform data entropy analysis based on the initial factoring data and the target sensitive information to obtain a first encryption method corresponding to the target sensitive information and a second encryption method corresponding to the initial factoring data; A data encryption module, which is used to encrypt the target sensitive information and the initial factoring data according to the first encryption method and the second encryption method to obtain a target encryption result; An intention recognition module, which is used to recognize the intention of the initial factoring data to obtain a target intention corresponding to the initial factoring data, and determine a target permission corresponding to the target encryption result according to the target intention; A data transmission module, which is used to transmit the target encryption result and the target permission to relevant nodes by using a blockchain, and obtain relevant permissions corresponding to the relevant nodes; A data screening module, which is used to screen the relevant nodes according to the target permission and the relevant permissions to obtain target nodes, so that the target object can obtain the target encryption result according to the target nodes.

Images