Blockchain-based electronic data notarization system and method

Abstract

This invention relates to the field of blockchain technology, specifically to a blockchain-based electronic data notarization system and method. The system includes a data preprocessing module, an originality verification module, a credibility verification module, and a result output module. The originality verification module compares the comparison hash value with the content hash value, generates an originality judgment result, and transmits it to the credibility verification module and the result output module. The credibility verification module calculates the comprehensive credibility of the verified data based on a credibility base score, a base weight, an authority assessment score, and an authority assessment weight, and transmits this result to the result output module. Through the design of the originality verification module and the credibility verification module, this invention can determine the authenticity of electronic data and analyze its credibility while ensuring the integrity and originality of the electronic data itself, thereby effectively improving the credibility and evidentiary value of the electronic data.

Description

Technical Field

[0001] This invention relates to the field of blockchain technology, and more specifically to a blockchain-based electronic data notarization system and method. Background Technology

[0002] In the current wave of digitalization, the widespread application of electronic data has made its security a key issue. Against this backdrop, blockchain-based electronic data notarization technology has emerged, providing a new approach to the preservation, storage, and verification of electronic data.

[0003] Many existing blockchain-based electronic data notarization systems have demonstrated unique advantages in practical applications. For example, the electronic data preservation evidence system of Tianjin Chenxin Notary Office uses blockchain technology, trusted timestamps, and hash value verification to build a legally recognized, convenient, and efficient electronic data preservation and judicial notarization service system. Through the application of blockchain technology, electronic data can be completely recorded and cannot be tampered with, thereby ensuring the integrity and originality of the electronic data, making it usable as proof for users to handle various matters.

[0004] While existing electronic data notarization technologies can ensure that electronic data is completely recorded and cannot be tampered with, this approach can only guarantee the integrity and originality of the electronic data itself, but cannot judge the authenticity of the electronic data itself, resulting in insufficient credibility and evidentiary value of the electronic data. Therefore, it is necessary to propose a blockchain-based electronic data notarization system and method that can judge the authenticity of electronic data and analyze its credibility while ensuring the integrity and originality of the electronic data itself. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a blockchain-based electronic data notarization system and method. Through the design of an originality verification module and a credibility verification module, it is possible to determine the authenticity of electronic data and analyze its credibility while ensuring the integrity and originality of the electronic data itself, thereby effectively improving the credibility and evidentiary value of the electronic data.

[0006] To achieve the above objectives, the technical solution of the present invention is as follows: a blockchain-based electronic data notarization system, comprising a data preprocessing module, an originality verification module, a credibility verification module, and a result output module.

[0007] The data preprocessing module is used to collect multi-source data and user identification data, analyze the content of multi-source data based on semantic recognition algorithms, extract the publishing entities and their types from the multi-source data, generate content hash values ​​based on the multi-source data content, and transmit the content hash values, publishing entities and their types to the original authenticity verification module; it also generates verification hash values ​​based on user identification data and transmits the verification hash values ​​to the credibility verification module.

[0008] The originality verification module receives content hash values, publishing entities, and their types from multi-source data, associates and stores them, and generates a content library based on blockchain distributed node storage. The originality verification module also allows users to input verification data and content hash values. Based on the user-input verification data, the module generates a comparison hash value, compares this comparison hash value with the content hash value, generates an originality judgment result, and transmits it to the trustworthiness verification module and the result output module. If the comparison hash value and the content hash value are completely identical, the verification data is determined to have not been modified and is considered original data; if they are not completely identical, the verification data is determined to have been modified and is considered non-original data.

[0009] The credibility verification module analyzes the content of the verification data based on semantic recognition algorithms, extracts the publishing entity and its type, retrieves historical multi-source data of the publishing entity from the content library based on the publishing entity, analyzes the correlation between the historical multi-source data and the current verification data based on semantic recognition algorithms and the type of the publishing entity, sets a basic credibility score for the verification data based on the correlation between the historical multi-source data and the current verification data, and sets a basic weight based on the type of the publishing entity; the correlation includes content correlation, data correlation and time correlation.

[0010] The credibility verification module is also used to connect to authoritative nodes and identify the types of authoritative nodes, perform credibility verification based on authoritative nodes, obtain an authority assessment score based on the credibility verification results, and set authority assessment weights based on the type of authoritative node.

[0011] The credibility verification module is also used to calculate the overall credibility of the verification data based on the credibility base score, base weight, authority assessment score, and authority assessment weight, and transmit it to the result output module.

[0012] The credibility verification module is also used to receive the originality judgment result, identify the originality of the verification data based on the originality judgment result, and if the verification data is not original data, credibility verification and credibility base score setting will not be performed.

[0013] The results output module is used to receive the originality judgment results and overall credibility of the verification data, and output the originality judgment results and overall credibility of the verification data to the user terminal for display.

[0014] Furthermore, the credibility verification process is as follows: the credibility verification module transmits the verification data to the authoritative node, the authoritative node evaluates the credibility of the verification data, generates an evaluation score and feeds it back to the credibility verification module, and calculates the authoritative evaluation score based on the weight of the authoritative node and the evaluation score.

[0015] Furthermore, the credibility verification module is also used to receive verification hash values ​​and generate a verification library, which is stored based on distributed nodes of the blockchain. The credibility verification module is used for users to input verification hash values ​​and content hash values. The credibility verification module judges whether the verification hash values ​​and content hash values ​​input by the user are correct and correspond based on the content library and the verification library. If both the hash values ​​and content hash values ​​are input correctly and correspond, the content of the corresponding multi-source data is displayed.

[0016] Furthermore, the types of entities that can publish information include government agencies, listed companies, ordinary enterprises, social organizations, and individual users.

[0017] Furthermore, the types of authoritative nodes include notary offices, law firms, and industry associations.

[0018] Furthermore, user identification data includes one or more of the following: user fingerprints, user faces, and user irises. User identification data also includes the time point at which the multi-source data was stored.

[0019] Furthermore, the formula for calculating the basic credibility score is as follows: X=(G1+G2+G3)·K0 (1).

[0020] Where X is the basic credibility score, G1 is the content relevance score, G2 is the data relevance score, G3 is the time relevance score, and K0 is the basic weight, with a value ranging from 0.1 to 1. The semantic recognition algorithm includes the BERT pre-trained model, which is used to extract entity information, relational information, and time information from the validation data and multi-source data and compare them. Based on the comparison results, the content relevance score, data relevance score, and time relevance score are output. If the publishing entity has no historical multi-source data, the basic credibility score is set to 50 points by default.

[0021] Furthermore, the formula for calculating the authoritative assessment score is as follows: Y=(A×K1+B×K2+C×K3) / (K1+K2+K3) (2).

[0022] Where Y is the authoritative assessment score, A is the notary office's assessment score, K1 is the notary office's authoritative assessment weight, B is the law firm's assessment score, K2 is the law firm's authoritative assessment weight, C is the industry association's assessment score, and K3 is the industry association's authoritative assessment weight; the values ​​of K1, K2, and K3 range from 0.1 to 1. The formula for calculating overall credibility is as follows: W=[(X+Y) / 2]% (3)。

[0023] Where W represents the overall credibility of the verification data.

[0024] Furthermore, after collecting multi-source data and user identification data, analyzing the content of multi-source data based on semantic recognition algorithms, and extracting the publishing entities and types of multi-source data, the data preprocessing module is also used to perform rationality verification on the content of multi-source data. Only after the verification passes will a content hash value be generated based on the content of multi-source data and subsequent data transmission be completed. If the verification fails, no content hash value will be generated. The rationality verification includes feature vector similarity verification and logical rule compliance verification. Feature vector similarity verification includes: the data preprocessing module processes multi-source data content into semantic logic blocks based on a semantic recognition algorithm to obtain several semantic data blocks, extracts time information and event development information from each semantic data block, and constructs a high-dimensional vector space based on time logic and event development logic. Map each semantic data block to a high-dimensional vector space. eigenvectors within ,in, This is the sequence number of the semantic data block. , Given the total number of semantic data blocks, calculate the semantic similarity between any two feature vectors. Furthermore, by using a preset similarity threshold, it is determined whether there are potential logical contradictions in the content of multi-source data. The formulas for calculating semantic similarity and average similarity are as follows: ; ; The data preprocessing module presets a similarity threshold. , ,like If so, it is determined that there are no potential logical contradictions in the content of the multi-source data, and the rationality check passes; like If the multi-source data content has potential logical contradictions, then all data that satisfy the condition will be extracted. contradictory semantic data block group Perform logical rule compliance checks on this set of data blocks, extract time information and event development information from contradictory semantic data blocks, and calculate the time logic compliance coefficient. Coefficient of conformity with the logic of the development of things The formula for calculating the comprehensive logic compliance is:

[0025] The data preprocessing module presets a comprehensive logic compliance threshold. , ,like If the contradiction is deemed to be a reasonable situation involving error correction or content supplementation of the multi-source data itself, the reasonableness check passes; if If the result is not clear, it is determined to be a logical contradiction caused by data omission, deliberate modification or material omission. The rationality check fails and the contradiction type and the location of the corresponding semantic data block are fed back to the user. in, For the first Semantic data blocks in a high-dimensional vector space The feature vector mapped in the text consists of core feature dimensions of time information and event development information, and serves as a unique vector identifier for the entire text. For the first The and the first The semantic similarity between feature vectors of semantic data blocks, with a value range of... The closer the value is to 1, the higher the semantic and logical correlation between the two. for and The vector dot product represents the degree of correlation between two feature vectors in the same direction; for The vector magnitude represents the first... The dimensional feature strength of the feature vector of each semantic data block; To ensure compatibility with different types of entities, government agencies Listed companies ordinary enterprises social organizations Individual users It adapts to the content expression logic characteristics of different subjects; The average semantic similarity among the feature vectors of all semantic data blocks, with a value range of... It represents the overall semantic and logical consistency of multi-source data content; From The number of combinations of two elements selected from each semantic data block is used as the normalized denominator for calculating the average similarity. The first-level similarity check threshold is a preset fixed value used as a criterion for defining potential logical contradictions. The temporal logical compliance coefficient for contradictory semantic data blocks, with a value range of... The closer the value is to 1, the more the temporal logic, such as the order of time and the time span between data blocks, conforms to objective laws. The coefficient for logical consistency of the development of contradictory semantic data blocks, with a value range of... The closer the value is to 1, the more the logic of cause and effect, development stages and other aspects between data blocks conforms to objective laws. The weight of the time logic compliance coefficient, with a value range of... This indicates the importance of time logic in comprehensive judgment; The weight of the coefficient for the logical consistency of the development of things, and its range of values. And satisfy This indicates the importance of the logic of the development of things in comprehensive judgment; The comprehensive logical compliance coefficient for contradictory semantic data blocks, with a value range of... This characterizes whether the logical contradiction in a contradictory data block is a reasonable situation. The threshold for logical compliance verification is a preset fixed value used to define the criteria for judging reasonable logical situations and obvious logical contradictions.

[0026] The above approach has the following beneficial effects: 1. Existing electronic data notarization technologies often rely solely on blockchain technology and hash values ​​to verify whether electronic data has been tampered with, and to determine the integrity and originality of the data. However, they cannot assess the authenticity of the document content itself. For example, a forged original document, even if it has not been tampered with, lacks authenticity and credibility. This system, while verifying originality, analyzes the content relevance, data relevance, and temporal relevance of the electronic data. Combining this with the analysis from various authoritative nodes, it comprehensively assesses the authenticity and credibility of the electronic data, thereby providing a data-driven demonstration of the reliability and evidentiary value of the electronic data.

[0027] 2. In existing electronic data notarization technologies, users often rely solely on notary offices, law firms, and industry associations to assess the credibility of electronic data. This assessment process carries the risk of institutional negligence or internal data tampering. Compared to existing technologies, this system calculates the overall credibility of electronic data by combining a basic credibility score with an authoritative assessment score, resulting in a more comprehensive assessment process. Furthermore, it adjusts the basic weight of the credibility score based on the type of the issuing entity and the authoritative assessment weight based on the type of authoritative node, thereby deriving the overall credibility of the electronic data. This reduces the impact of institutional negligence or data tampering on the assessment results and improves their accuracy.

[0028] 3. In existing electronic data notarization technologies, hash values ​​lack a binding relationship with users. Anyone can query and notarize target electronic data by inputting the target hash value, resulting in a lack of protection for user privacy. This system, through the separation of content hash value and verification hash value, allows anyone to query the integrity and originality of target electronic data through the content hash value, but not to view the content of the electronic data. Only the user or a user authorized by the user to verify the hash value can view the content of the electronic data by inputting both the correct and corresponding content hash value and verification hash value. This strengthens the binding relationship between electronic data and users, solves the problem of the separation between hash value and user identity in existing technologies, and also ensures the integrity and originality of electronic data.

[0029] Furthermore, the blockchain-based electronic data notarization method includes the following steps: S1, Data Input: Input multi-source data and user identification data into the data preprocessing module to obtain content hash value and verification hash value.

[0030] S2, Data Originality Verification: Input the verification data into the originality verification module, obtain the originality judgment result, and determine whether the verification data has been modified based on the originality judgment result.

[0031] S3, Data Credibility Verification: If the verification data has not been modified, the verification data is input into the credibility verification module to obtain the overall credibility of the verification data.

[0032] The above approach has the following beneficial effects: Existing electronic data notarization methods can only determine whether electronic data has been modified. Even if the electronic data has not been modified, it does not mean that it is authentic and credible. Therefore, existing electronic data notarization methods cannot determine the authenticity and credibility of electronic data. This method, by verifying the authenticity of electronic data while judging its credibility, reflects the authenticity and evidentiary effect of electronic data.

[0033] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the structure of the blockchain-based electronic data notarization system of the present invention.

[0035] Figure 2 This is a flowchart illustrating the originality determination in the blockchain-based electronic data notarization system of the present invention.

[0036] Figure 3 This is a flowchart illustrating the comprehensive credibility calculation in the blockchain-based electronic data notarization system of this invention.

[0037] Figure 4 This is a schematic diagram illustrating the steps of the blockchain-based electronic data notarization method of the present invention. Detailed Implementation

[0038] The following detailed description illustrates the specific implementation method: Example 1: As attached Figure 1 As shown: A blockchain-based electronic data notarization system includes a data preprocessing module, an originality verification module, a credibility verification module, and a result output module.

[0039] The specific functions of each module are as follows: The data preprocessing module is used to collect multi-source data (documents, images, and audio, etc.) and user identification data (including one or more of user fingerprints, user faces, and user irises, as well as the time nodes when the multi-source data was stored). Based on semantic recognition algorithms, it analyzes the content of the multi-source data, extracts the publishing entity and its type, generates content hash values ​​based on the multi-source data content, and transmits the content hash values, publishing entities, and their types to the original authenticity verification module. Based on the user identification data, it generates verification hash values ​​and transmits the verification hash values ​​to the credibility verification module.

[0040] Specifically, assuming user A uploads an electronic contract to the data preprocessing module and registers their fingerprint on May 20, 2024, the data preprocessing module will generate a content hash value and a verification hash value for the electronic contract. For example, the content hash value is a3b7c937X, and the verification hash value is SHA-256. Both the content hash value and the verification hash value are unique.

[0041] like Figure 2As shown, the originality verification module receives the content hash value, publishing entity, and type of multi-source data, associates and stores them, and generates a content library based on blockchain distributed node storage. The originality verification module also allows users to input verification data and content hash values. Based on the user-input verification data, the module generates a comparison hash value, compares it with the content hash value, generates an originality judgment result, and transmits it to the trustworthiness verification module and the result output module. If the comparison hash value and the content hash value are completely identical, the verification data is determined to have not been modified and is considered original data. If the comparison hash value and the content hash value are not completely identical, the verification data is determined to have been modified and is considered non-original data.

[0042] Specifically, suppose user A needs to prove to a business partner that an electronic contract has not been modified. User A inputs the current electronic contract into the originality verification module. The originality verification module will output a comparison hash value based on the electronic contract uploaded by user A. Suppose the comparison hash value is: a3b7c937X. The comparison hash value is consistent with the content hash value. Therefore, the electronic contract submitted by user A has not been modified and is original data.

[0043] like Figure 3 As shown, the credibility verification module analyzes the content of the verification data based on semantic recognition algorithms, extracts the publishing entity and its type, retrieves historical multi-source data of the publishing entity from the content library based on the publishing entity, analyzes the correlation between the historical multi-source data and the current verification data based on semantic recognition algorithms and the type of the publishing entity, sets the basic credibility score of the verification data according to the correlation between the historical multi-source data and the current verification data, and sets the basic weight based on the type of the publishing entity; the correlation includes content correlation, data correlation and time correlation.

[0044] The types of entities that publish information include government agencies, listed companies, ordinary enterprises, social organizations, and individual users (in this embodiment, the basic weights of government agencies, listed companies, ordinary enterprises, social organizations, and individual users are 1, 0.8, 0.7, and 0.6, respectively).

[0045] The formula for calculating the basic credibility score is as follows: X=(G1+G2+G3)·K0 (1).

[0046] Where X is the basic credibility score, G1 is the content relevance score, G2 is the data relevance score, G3 is the time relevance score, and K0 is the basic weight, with a value range of 0.1-1. The semantic recognition algorithm includes the BERT pre-trained model, which is used to extract entity information, relational information, and time information from the validation data and multi-source data and compare them. Based on the comparison results, it outputs the content relevance score (maximum 50 points), the data relevance score (maximum 25 points), and the time relevance score (maximum 25 points). If the publishing entity has no historical multi-source data, the basic credibility score is set to 50 points by default.

[0047] Specifically, after User A's electronic contract passes the originality verification, the partner requests further confirmation of the credibility of the contract content. The credibility verification module analyzes the contract content based on the semantic recognition algorithm (BERT pre-trained model): extracts the publishing entity, "User A (ordinary enterprise, basic weight is 0.6)", and retrieves User A's historical multi-source data through the content library. It is found that User A's three previous electronic contracts were all commercial cooperation contracts, with some content related but some differences, and the content relevance score is 40 points. The names and unified social credit codes of the companies of both parties in the electronic contract are consistent with the records of the National Enterprise Credit Information Publicity System, and there is no business abnormality, so the data relevance score is 25 points. The contract signing time (May 18, 2024) is earlier than the evidence storage time (May 20, 2024), and the time logic is reasonable, so the time relevance score is 25 points. According to formula (1), User A's credibility base score K=63.

[0048] The credibility verification module also connects to and identifies authoritative nodes, performs credibility verification based on the authoritative nodes, obtains an authority assessment score based on the credibility verification results, and sets authority assessment weights based on the authority node type. The credibility verification process is as follows: the credibility verification module transmits verification data to the authoritative node, the authoritative node assesses the credibility of the verification data, generates an assessment score, and feeds it back to the credibility verification module. Based on the authority node's weight and the assessment score, the authority assessment score is calculated. The credibility verification module also receives the originality judgment result, identifies the originality of the verification data based on the originality judgment result, and if the verification data is not original data, credibility verification and the setting of the basic credibility score are not performed.

[0049] The types of authoritative nodes include notary offices, law firms, and industry associations (the authority assessment weights for notary offices, law firms, and industry associations are 0.5, 0.3, and 0.2, respectively).

[0050] The formula for calculating the authoritative assessment score is as follows: Y=(A×K1+B×K2+C×K3) / (K1+K2+K3) (2).

[0051] Where Y is the authoritative assessment score, A is the assessment score of the notary office, K1 is the authoritative assessment weight of the notary office, B is the assessment score of the law firm, K2 is the authoritative assessment weight of the law firm, C is the assessment score of the industry association, and K3 is the authoritative assessment weight of the industry association; the values ​​of K1, K2 and K3 range from 0.1 to 1.

[0052] Specifically, assuming that the notary office, law firm and industry association give evaluation scores of 90, 85 and 88 respectively for user A's electronic contract, according to formula (2), the authoritative evaluation score of the electronic contract is Y=88.1, and finally Y=88.

[0053] The credibility verification module is also used to calculate the overall credibility of the verification data based on the credibility base score, base weight, authority assessment score, and authority assessment weight, and transmit it to the result output module.

[0054] The formula for calculating overall credibility is as follows: W=[(X+Y) / 2]% (3)。

[0055] Where W represents the overall credibility of the verification data.

[0056] Specifically, taking the above-mentioned electronic contract as an example, according to formula (3), the overall credibility of the electronic contract is W=75.5%.

[0057] The results output module is used to receive the originality judgment results and overall credibility of the verification data, and output the originality judgment results and overall credibility of the verification data to the user terminal for display.

[0058] Specifically, the output module will display the originality assessment result and overall credibility of the electronic contract: the electronic data has not been modified and is original data; the overall credibility of the electronic data is 75.5%, which is highly credible.

[0059] The credibility verification module is also used to receive verification hash values ​​and generate a verification library, which is stored based on distributed nodes in the blockchain. The credibility verification module is used for users to input verification hash values ​​and content hash values. The credibility verification module judges whether the verification hash values ​​and content hash values ​​input by the user are correct and correspond based on the content library and the verification library. If both the hash values ​​and content hash values ​​are input correctly and correspond, the content of the corresponding multi-source data is displayed.

[0060] Specifically, suppose user A needs to view the content of an electronic contract. By inputting the verification hash value (SHA-256) and the content hash value (a3b7c937X) into the credibility verification module, if the two are correct and correspond, the verification is successful, and the credibility verification module will then display the content of the electronic contract.

[0061] Existing electronic data notarization technologies often rely solely on blockchain technology and hash values ​​to verify whether electronic data has been tampered with, determining its integrity and originality. However, they cannot assess the authenticity of the document's content itself. For example, a forged original document, even without tampering, lacks authenticity and credibility. This system, while verifying originality, analyzes the content relevance, data relevance, and temporal relevance of the electronic data. Combining this with analyses from various authoritative nodes, it comprehensively evaluates the authenticity and credibility of the electronic data, thereby providing a data-driven demonstration of its reliability and evidentiary value.

[0062] Example 2: like Figure 4 As shown, unlike the above embodiments, the blockchain-based electronic data notarization method includes the following steps: S1, Data Input: Input multi-source data and user identification data into the data preprocessing module to obtain content hash value and verification hash value.

[0063] S2, Data Originality Verification: Input the verification data into the originality verification module, obtain the originality judgment result, and determine whether the verification data has been modified based on the originality judgment result.

[0064] S3, Data Credibility Verification: If the verification data has not been modified, the verification data is input into the credibility verification module to obtain the overall credibility of the verification data.

[0065] I. Experimental Preparation Experimental subjects: 100 electronic files (including contracts, certificates, pictures and audio) were selected, of which 50 were original files that had not been tampered with (Group A), and 50 were files that had been partially tampered with or forged but whose hash values ​​did not conflict (Group B).

[0066] Experimental environment: The above 100 electronic documents were processed using this method and system.

[0067] Evaluation indicators: Originality verification accuracy (the accuracy of the consistency judgment based on hash value comparison); Credibility assessment score distribution (reasonableness of overall credibility W); System processing time (average time from data input to result output).

[0068] II. Experimental Procedure Step 1: Perform data preprocessing on each group of files to generate content hash values ​​and verification hash values, and store them in the blockchain.

[0069] Step 2: Verify the originality of all 100 documents and record the results.

[0070] Step 3: Verify the credibility of the files that have passed the originality verification, extract the type of the publishing entity and the correlation of historical data, and call the authoritative node for evaluation.

[0071] Step 4: Calculate the accuracy rate of authenticity judgment and the accuracy rate of credibility assessment. (Group A contains 25 genuine documents and 25 forged documents; if the system outputs a credibility rate > 50%, the system determines that the document is a genuine document; if the system outputs a credibility rate ≤ 50%, the system determines that the document is a forged document.) III. Experimental Results Table 1 Statistical table of analysis results

[0072] Table 2 Statistical Table for Authenticity Judgment

[0073] IV. Experimental Conclusions Originality verification accuracy: According to Table 1, Group A (50 untampered files) was correctly identified as original data, while Group B (50 tampered files) was identified as non-original data. The accuracy rate of originality judgment was 100%, and the originality judgment function of this method is stable and effective.

[0074] Credibility assessment results: According to Tables 1 and 2, the reliability assessment accuracy of this method was 98% in this experiment. All real documents were correctly identified, but one forged document was identified as a real document. This may be because a very small number of forged documents were too realistic, which affected the accuracy of the assessment. However, the reliability assessment accuracy of this method is still as high as 98%, and the reliability assessment function of this method is stable and effective.

[0075] In summary, this system not only reliably verifies the originality of electronic data, but also achieves a quantitative assessment of the authenticity of data content through multi-dimensional correlation analysis and the intervention of authoritative nodes, significantly improving the comprehensiveness and credibility of electronic data notarization technology.

[0076] Example 3 Compared to Example 1, the only difference is that, after collecting multi-source data and user identification data, analyzing the multi-source data content based on semantic recognition algorithms, and extracting the publishing entities and types of the multi-source data, the data preprocessing module adds a two-level verification process for the rationality of the multi-source data content. Only after the verification passes will a content hash value be generated based on the multi-source data content and subsequent data transmission to the original rationality verification module be completed. If the verification fails, no content hash value is generated, and the contradiction type and the location of the corresponding semantic data block are fed back to the user. The rationality verification process is explained in conjunction with actual application scenarios as follows: In this embodiment, user B, a publisher of ordinary enterprise type, is selected as the evidence storage subject. On August 10, 2024, user B uploaded a supplementary business cooperation agreement to the data preprocessing module as multi-source data, and at the same time entered facial information as user identification data. The supplementary agreement is a document in which user B and the partner corrected the terms of the previous main contract and corrected some typographical errors. The data preprocessing module performs a two-level reasonableness check on the supplementary agreement.

[0077] The data preprocessing module, based on semantic recognition algorithms (such as the BERT pre-trained model), divides the supplementary business cooperation agreement into semantic blocks according to semantic logic. These blocks are categorized into five semantic data blocks: "Main Agreement Clauses," "Supplementary Clauses," "Payment Clauses," "Performance Period Clauses," and "Breach of Contract Clauses." These are sequentially labeled as semantic data block 1, semantic data block 2, semantic data block 3, semantic data block 4, and semantic data block 5. Subsequently, time information (such as performance period and payment time) and event development information (such as contract performance process and default handling process) are extracted from each semantic data block, and a high-dimensional vector space is constructed based on time logic and event development logic. The five semantic data blocks are mapped to high-dimensional vector spaces. eigenvectors within , , , , Each feature vector is composed of core feature dimensions such as time node, performance stage, and rights and obligations relationship.

[0078] First, feature vector similarity verification is performed. Since the publishing entity in this embodiment is a regular enterprise, based on the definition of the publishing entity type adaptation coefficient, this embodiment takes... The data preprocessing module presets a first-level similarity threshold. Then according to the formula Calculate the semantic similarity between any two feature vectors Then through the formula Calculate the average semantic similarity among all feature vectors. ; The semantic similarity between semantic data block 2 (correction clause) and semantic data block 1 (protocol body clause) was calculated. Semantic similarity with semantic data block 4 (performance period clause) The semantic similarity between the remaining feature vectors is all above 0.75. After substituting them into the average similarity formula, the overall average semantic similarity is obtained. ,because The data preprocessing module determines that the supplementary protocol has a potential logical contradiction and extracts the data that satisfies this contradiction. contradictory semantic data block group , The second-level logical rule compliance check is performed on the two sets of contradictory data blocks.

[0079] Then, logical rule compliance verification is performed, and the data preprocessing module extracts contradictory semantic data block groups. , Calculate the time logic conformity coefficient for the time information and the information on the development of events. Coefficient of conformity with the logic of the development of things In this embodiment, a preset time logic compliance coefficient weight is used. Weight of the consistency coefficient of the development logic of things (satisfy (Preset second-level comprehensive logic compliance threshold) Through formula Calculate the synthesis logic compliance coefficient ; Among them, the time logic compliance coefficient The coefficient of logical conformity to the development of things is derived by measuring the matching of the temporal order and time span between contradictory data blocks with objective laws. This is derived from the quantification of the matching between the causal relationships, development stages, and objective laws governing the performance of commercial contracts among contradictory data blocks.

[0080] Specifically, the time logic compliance coefficient The range of values ​​is Its quantification focuses on the degree of matching between the time elements of contradictory data blocks and objective time patterns, first extracting the core time nodes within the contradictory data blocks. ( (Based on the time node number), then matching degree based on the chronological order. Reasonableness and matching degree of time span The two dimensions are quantified, and the final result is obtained through weighted summation. The quantification formula is:

[0081] in: The time sequence matching weight is set to a value of [value]. , which is a preset fixed value, because the objectivity of the order of time is the core element of the time logic; The weight for the reasonableness of the time span matching is set to a value of [value]. , is a preset fixed value, satisfying ; The time sequence matching degree, with values ​​ranging from 1 to 2. If the core time nodes between contradictory data blocks completely conform to objective chronological laws, then take... Partially meets the requirements. This is completely inconsistent with the requirement to take ; The value is used to determine the reasonableness of the time span. If the time span between contradictory data blocks matches the objective execution cycle of the corresponding transaction, then take... Slight deviation Serious deviation .

[0082] For example, regarding contradictory data block groups The core time nodes of semantic data block 2 (supplementary clause) are August 10, 2024 (the date of signing the supplementary agreement) and July 2, 2024 (the start date of the supplementary clause for the performance period of the main contract). The original time node recorded in semantic data block 4 (performance period clause) is July 1, 2024 (the original start date of the performance period of the main contract). Time sequence matching degree The revised performance period commencement date of July 2, 2024, falls on the evening of July 1, 2024, the signing date of the main contract, and is earlier than the signing date of the supplementary agreement, August 10, 2024. This fully conforms to the objective temporal pattern of performance period revisions in commercial contracts. Therefore... ; Reasonableness and matching degree of time span The corrected performance period differs from the original performance period by only one day, a minor discrepancy caused by a clerical error. This has no significant conflict with the performance cycle of a commercial contract and is entirely consistent with objective reasonableness. ; Then the contradictory data block group .

[0083] Targeting contradictory data block groups Semantic data block 2 only corrects typos in the partner names of semantic data block 1, without adding any new time nodes. The time element perfectly matches objective laws, so it is directly taken. ; In this embodiment, two sets of contradictory data blocks are selected. The arithmetic mean of these is the final time logic compliance coefficient, i.e. .

[0084] Coefficient of conformity of the logic of the development of things The range of values ​​is Its quantification focuses on the matching degree between the development elements of things in contradictory data blocks and the objective laws of commercial contract performance. First, it extracts the core development nodes of things within the contradictory data blocks. ( (as the sequence number of the development nodes of things), and then from the degree of matching of causal relationships. Development stage matching degree The two dimensions are quantified, and the final result is obtained through weighted summation. The quantification formula is:

[0085] in: The weight of the causal relationship matching degree, with a value of , which is a preset fixed value, because the cause-and-effect relationship is the core element of the logic of the development of things; The matching degree weight for the development stage is set to a value of [value]. , is a preset fixed value, satisfying ; The degree of causal relationship matching, with values ​​ranging from 1 to 2. If the causal relationship between contradictory data blocks fully conforms to objective business laws, then take... Partially meets the requirements. This is completely inconsistent with the requirement to take ; To determine the matching degree at the development stage, the value is [value to be filled in]. If the development stage of things between contradictory data blocks matches the performance stage of an objective commercial contract, then take... Slight deviation Serious deviation .

[0086] For example, regarding contradictory data block groups The core development node of semantic data block 2 (supplementary clause) is "supplementing the performance period of the main contract → the supplementary clause shall take effect from the date of signing the supplementary agreement", and the core development node of semantic data block 4 (performance period clause) is "the performance period of the main contract shall be executed from the original start date". Causal relationship matching degree The supplementary agreement's correction of the main contract's performance period was a rectification of a clerical error in the original performance period. The causal relationship of the supplementary clause replacing the original performance period after it takes effect is entirely consistent with the objective causal laws governing the correction of commercial contracts. ; Development stage matching degree Since the main contract is still in the incomplete stage, signing a supplementary agreement to adjust the performance period at this time perfectly matches the objective developmental pattern of commercial contracts: "clauses can be amended before / during performance, but there is no need for amendments after performance is completed." ; Substituting into the formula yields the contradictory data block group. .

[0087] Targeting contradictory data block groups The core development node of semantic data block 2 is "correcting the typo in the name of the main contract partner → the corrected main clauses are the basis for the effective execution of the contract", and the core development node of semantic data block 1 is "agreeing on the main information of the partner → as the basic element of contract performance". Causal relationship matching degree Because a misspelling of the main contract's name could lead to an error in identifying the party responsible for contract performance, correcting the name is essential to ensure the proper performance of the contract. This causal relationship is entirely consistent with the objective laws of commercial contracts. ; Development stage matching degree Since the main contract was signed but not fully performed, and the supplementation of the subject information is at a reasonable stage of contract performance, therefore... ; Substituting into the formula yields the contradictory data block group. ; In this embodiment, two sets of contradictory data blocks are selected. The arithmetic mean is the coefficient of logical consistency of the final development of things, that is... .

[0088] For example, upon verification, Semantic Data Block 2 (Correction Clause) corrected a typo in the subject name of Semantic Data Block 1 and reasonably corrected the performance period of Semantic Data Block 4. These are common and legitimate corrections and content amendments in commercial contracts, and are not contradictions caused by data omissions, deliberate modifications, or missing materials. If the temporal logic compliance coefficient of this contradictory data block group is considered... The coefficient of conformity to the logic of the development of things is quantified as 0.8. The quantization is 0.7. If we substitute it into the formula, we get the comprehensive logic compliance coefficient. ,because The data preprocessing module determined that the potential contradiction was a reasonable situation, such as error correction and content supplementation of the multi-source data itself, and the rationality verification passed.

[0089] Since the rationality of the supplementary business cooperation agreement passed both levels of verification, the data preprocessing module generates a unique content hash value based on the content of the supplementary agreement. At the same time, the content hash value, the publishing entity (user B) and its type (ordinary enterprise) are transmitted to the originality verification module. Based on the facial information entered by user B, a verification hash value is generated and transmitted to the credibility verification module. The subsequent originality verification, credibility verification and result output process is the same as in Example 1, and will not be repeated here.

[0090] To further disclose the implementation scenarios of this technical solution, this embodiment also supplements the description of application scenarios where the rationality check fails: If, in the supplementary business cooperation agreement uploaded by user B to the data preprocessing module, semantic data block 2 (correction clause) and semantic data block 3 (payment clause) have a significant temporal logic conflict (e.g., the corrected payment time is earlier than the contract signing time), and the causal logic of events is reversed (e.g., the supplementary agreement is signed after the agreement is fulfilled), then the first-level feature vector similarity check will result in... After extracting the contradictory data block group, perform a second-level logical rule compliance check. If quantification yields... , If the comprehensive logic compliance coefficient ,because The data preprocessing module determined that the contradiction was a clear logical contradiction caused by data omission, deliberate modification, or material omission. The rationality check failed, no content hash value was generated, and the specific contradiction type was fed back to the user as "time logic conflict + reversed cause and effect of events". At the same time, the contradictory data blocks were marked as "semantic data block 2 and semantic data block 3" to facilitate user B to correct data omissions, supplement missing materials, or re-upload real and valid multi-source data in a timely manner.

[0091] For example, if there is a clear contradiction between semantic data block 2 (correction clause) and semantic data block 3 (price payment clause) in the supplementary business cooperation agreement uploaded by user B to the data preprocessing module, the core time node of semantic data block 2 is extracted as June 1, 2024 (the corrected price payment start date), while the original time node recorded in semantic data block 3 is July 1, 2024 (the main contract price payment start date), and the main contract signing time is July 1, 2024; the core development node of semantic data block 2 is "the main contract is signed after the price payment is completed", while the core development node of semantic data block 3 is "the price payment is initiated after the main contract is signed".

[0092] Time logic compliance coefficient The revised payment start date of June 1, 2024, is earlier than the main contract signing date of July 1, 2024, which completely violates the objective laws of time sequence. This timeframe deviation of one month severely conflicts with the objective cycle of commercial contract payment processes. Substituting into the formula, we get ; Coefficient of conformity of the logic of the development of things Quantitatively, the objective causal relationship of "signing the main contract after payment" completely contradicts the commercial contract principle of "signing the contract first and then fulfilling the payment obligation." The payment occurred before the contract was signed, which is completely inconsistent with the "signing-performance-settlement" development stage of a commercial contract. Substituting into the formula, we get ; Calculation of the overall logic compliance coefficient: ,because The data preprocessing module determined that the contradiction was a clear logical contradiction caused by data omission, deliberate modification, or material omission. The rationality check failed, no content hash value was generated, and the specific contradiction type was fed back to the user as "time sequence conflict + reversed cause and effect of events". At the same time, the contradictory data blocks were marked as "semantic data block 2 and semantic data block 3" to facilitate user B to correct data omissions, supplement missing materials, or re-upload real and valid multi-source data in a timely manner.

[0093] In this embodiment, the data preprocessing module achieves accurate identification of multi-source data content through the aforementioned two-level rationality verification process. This avoids misjudging legitimate error correction and supplementary data as invalid data, and filters out false or invalid data caused by omissions, deliberate modifications, or missing materials from the source of evidence storage. This effectively improves the quality of blockchain evidence storage data, while reducing the amount of invalid data stored on the blockchain distributed nodes, lowering the computing power consumption for node data synchronization and retrieval, and making subsequent originality verification and credibility verification more accurate and effective.

[0094] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A blockchain-based electronic data notarization system, characterized in that: It includes a data preprocessing module, an originality verification module, a credibility verification module, and a result output module; The data preprocessing module is used to collect multi-source data and user identification data, analyze the content of multi-source data based on semantic recognition algorithms, extract the publishing entities and their types from the multi-source data, generate content hash values ​​based on the content of multi-source data, and transmit the content hash values, publishing entities and their types from the multi-source data to the original authenticity verification module. A verification hash value is generated based on user identification data, and the verification hash value is transmitted to the trustworthiness verification module. The original authenticity verification module is used to receive the content hash value, publishing entity and its type from multiple sources, and store them in association to generate a content library, which is stored based on blockchain distributed nodes. The originality verification module is also used for users to input verification data and content hash values. The originality verification module generates a comparison hash value based on the user-input verification data, compares the comparison hash value with the content hash value, generates an originality judgment result, and transmits it to the credibility verification module and the result output module. If the comparison hash value and the content hash value are completely consistent, it is determined that the verification data has not been modified and is original data; if the comparison hash value and the content hash value are not completely consistent, it is determined that the verification data has been modified and is not original data. The credibility verification module analyzes the content of the verification data based on semantic recognition algorithms, extracts the publishing entity and its type, retrieves historical multi-source data of the publishing entity from the content library based on the publishing entity, and analyzes the correlation between the historical multi-source data and the current verification data based on semantic recognition algorithms and the type of the publishing entity. Based on the correlation between the historical multi-source data and the current verification data, a basic credibility score for the verification data is set; a basic weight is set based on the type of the publishing entity; the correlation includes content correlation, data correlation, and time correlation. The credibility verification module is also used to connect to authoritative nodes and identify the types of authoritative nodes, perform credibility verification based on the authoritative nodes, and obtain the authority assessment score based on the credibility verification results. Set authority evaluation weights based on the type of authority node; The credibility verification module is also used to calculate the overall credibility of the verification data based on the credibility base score, base weight, authority assessment score, and authority assessment weight, and transmit it to the result output module. The credibility verification module is also used to receive the originality judgment result, identify the originality of the verification data based on the originality judgment result, and if the verification data is not original data, credibility verification and credibility base score setting will not be performed. The results output module is used to receive the originality judgment results and overall credibility of the verification data, and output the originality judgment results and overall credibility of the verification data to the user terminal for display.

2. The blockchain-based electronic data notarization system according to claim 1, characterized in that, The credibility verification process is as follows: The credibility verification module transmits verification data to the authoritative node. The authoritative node evaluates the credibility of the verification data, generates an evaluation score, and feeds it back to the credibility verification module. Based on the weight of the authoritative node and the evaluation score, the authoritative evaluation score is calculated.

3. The blockchain-based electronic data notarization system according to claim 1, characterized in that, The credibility verification module is also used to receive verification hash values ​​and generate a verification library, which is stored based on distributed nodes in the blockchain. The credibility verification module is used for users to input verification hash values ​​and content hash values. The credibility verification module judges whether the verification hash values ​​and content hash values ​​input by the user are correct and correspond based on the content library and the verification library. If both the hash values ​​and content hash values ​​are input correctly and correspond, the content of the corresponding multi-source data is displayed.

4. The blockchain-based electronic data notarization system according to claim 1, characterized in that, The types of entities that publish information include government agencies, listed companies, ordinary enterprises, social organizations, and individual users.

5. The blockchain-based electronic data notarization system according to claim 1, characterized in that, Authoritative nodes include notary offices, law firms, and industry associations.

6. The blockchain-based electronic data notarization system according to claim 1, characterized in that, User identification data includes one or more of the following: user fingerprints, user faces, and user irises. User identification data also includes the time point at which the multi-source data was stored.

7. The blockchain-based electronic data notarization system according to claim 1, characterized in that, The formula for calculating the basic credibility score is as follows: X=(G1+G2+G3)·K0 (1); Where X is the basic credibility score, G1 is the content relevance score, G2 is the data relevance score, G3 is the time relevance score, and K0 is the basic weight, with a value range of 0.1-1. The semantic recognition algorithm includes the BERT pre-trained model, which is used to extract entity information, relational information and time information from the validation data and multi-source data and compare them. Based on the comparison results, it outputs content relevance score, data relevance score and time relevance score. If the publishing entity has no historical multi-source data, the basic credibility score is set to 50 points by default.

8. The blockchain-based electronic data notarization system according to claim 1, characterized in that, The formula for calculating the authoritative assessment score is as follows: Y=(A×K1+B×K2+C×K3) / (K1+K2+K3) (2); Where Y is the authoritative assessment score, A is the notary office's assessment score, K1 is the notary office's authoritative assessment weight, B is the law firm's assessment score, K2 is the law firm's authoritative assessment weight, C is the industry association's assessment score, and K3 is the industry association's authoritative assessment weight; the values ​​of K1, K2, and K3 range from 0.1 to 1. The formula for calculating overall credibility is as follows: W = [(X+Y) / 2]% (3); Where W represents the overall credibility of the verification data.

9. The blockchain-based electronic data notarization system according to claim 1, characterized in that, After collecting multi-source data and user identification data, analyzing the content of multi-source data based on semantic recognition algorithms, and extracting the publishing entities and types of multi-source data, the data preprocessing module is also used to perform rationality verification on the content of multi-source data. Only after the verification passes will a content hash value be generated based on the content of multi-source data and subsequent data transmission be completed. If the verification fails, no content hash value will be generated. The rationality verification includes feature vector similarity verification and logical rule compliance verification. Feature vector similarity verification includes: the data preprocessing module processes multi-source data content into semantic logic blocks based on a semantic recognition algorithm to obtain several semantic data blocks, extracts time information and event development information from each semantic data block, and constructs a high-dimensional vector space based on time logic and event development logic. Map each semantic data block to a high-dimensional vector space. eigenvectors within ,in, This is the sequence number of the semantic data block. , Given the total number of semantic data blocks, calculate the semantic similarity between any two feature vectors. Furthermore, by using a preset similarity threshold, it is determined whether there are potential logical contradictions in the content of multi-source data. The formulas for calculating semantic similarity and average similarity are as follows: ； ； The data preprocessing module presets a similarity threshold. , ,like If so, it is determined that there are no potential logical contradictions in the content of the multi-source data, and the rationality check passes; like If the multi-source data content has potential logical contradictions, then all data that satisfy the condition will be extracted. contradictory semantic data block group Perform logical rule compliance checks on this set of data blocks, extract time information and event development information from contradictory semantic data blocks, and calculate the time logic compliance coefficient. Coefficient of conformity with the logic of the development of things The formula for calculating the comprehensive logic compliance is: The data preprocessing module presets a comprehensive logic compliance threshold. , ,like If the contradiction is deemed to be a reasonable situation involving error correction or content supplementation of the multi-source data itself, the reasonableness check passes; if If the result is not clear, it is determined to be a logical contradiction caused by data omission, deliberate modification or material omission. The rationality check fails and the contradiction type and the location of the corresponding semantic data block are fed back to the user. in, For the first Semantic data blocks in a high-dimensional vector space The feature vector mapped in the text consists of core feature dimensions of time information and event development information, and serves as a unique vector identifier for the entire text. For the first The and the first The semantic similarity between feature vectors of semantic data blocks, with a value range of... The closer the value is to 1, the higher the semantic and logical correlation between the two. for and The vector dot product represents the degree of correlation between two feature vectors in the same direction; for The vector magnitude represents the first... The dimensional feature strength of the feature vector of each semantic data block; To ensure compatibility with different types of entities, government agencies Listed companies ordinary enterprises social organizations Individual users It adapts to the content expression logic characteristics of different subjects; The average semantic similarity among the feature vectors of all semantic data blocks, with a value range of... It represents the overall semantic and logical consistency of multi-source data content; From The number of combinations of two elements selected from each semantic data block is used as the normalized denominator for calculating the average similarity. The first-level similarity check threshold is a preset fixed value used as a criterion for defining potential logical contradictions. The temporal logical compliance coefficient for contradictory semantic data blocks, with a value range of... The closer the value is to 1, the more the temporal logic, such as the order of time and the time span between data blocks, conforms to objective laws. The coefficient for logical consistency of the development of contradictory semantic data blocks, with a value range of... The closer the value is to 1, the more the logic of cause and effect, development stages and other aspects between data blocks conforms to objective laws. The weight of the time logic compliance coefficient, with a value range of... This indicates the importance of time logic in comprehensive judgment; The weight of the coefficient for the logical consistency of the development of things, and its range of values. And satisfy This indicates the importance of the logic of the development of things in comprehensive judgment; The comprehensive logical compliance coefficient for contradictory semantic data blocks, with a value range of... This characterizes whether the logical contradiction in a contradictory data block is a reasonable situation. The threshold for logical compliance verification is a preset fixed value used to define the criteria for judging reasonable logical situations and obvious logical contradictions.

10. A blockchain-based electronic data notarization method, operating based on the blockchain-based electronic data notarization system described in any one of claims 1-9, characterized in that, Includes the following steps: S1, Data Input: Input multi-source data and user identification data into the data preprocessing module to obtain content hash value and verification hash value; S2, Data Originality Verification: Input the verification data into the originality verification module, obtain the originality judgment result, and determine whether the verification data has been modified based on the originality judgment result; S3, Data Credibility Verification: If the verification data has not been modified, the verification data is input into the credibility verification module to obtain the overall credibility of the verification data.

Images