A Method for the Whole Process Data Preservation of Electronic Signatures Based on Blockchain

Abstract

This application relates to the technical field of electronic signature and evidence preservation, and specifically relates to a method for preserving all-process data of electronic signatures based on blockchain. The method includes: collecting all historical transaction behavior data of each user before each transaction in the blockchain, including transaction users, transaction times, and text data of signed documents; determining the theme difference degree of signed documents for any two transactions; determining the transaction behavior anomaly coefficient for each user at each transaction; obtaining the malicious user confidence level for each user at each transaction; for each transaction, based on the number of transactions between each user and each user to be signed, combined with the malicious user confidence level of each user to be signed, obtaining the selectable coefficient of each user to be signed for each user at each transaction, screening the users to be signed, and using an encrypted signature algorithm for electronic signature. This application ensures the anonymity of the signer's identity and improves the privacy and security of both trading parties.

Description

Technical Field

[0001] This application relates to the technical field of electronic signature and evidence preservation, and specifically relates to a method for preserving all-process data of electronic signatures based on blockchain. Background Art

[0002] With the popularization of e-commerce and digital transactions, electronic contracts have become an indispensable part of modern business activities, and electronic signatures are essential elements for the effectiveness of electronic contracts. However, the traditional methods of signing and preserving electronic contracts have the risk of data being tampered with and lost, and cannot effectively guarantee the security and credibility of the signed data.

[0003] At present, with the continuous development of information technology, blockchain technology, as a decentralized and immutable distributed ledger technology, has gradually become an ideal choice for solving the security and credibility problems of electronic signatures due to its transparency, security, and immutability. However, in practical applications, since the evidence-preserving data on the blockchain is publicly transparent, blockchain participants maintain a common ledger, and each participant can view and verify the evidence-preserving data of other participants, which increases the risk of user privacy data. Although the traditional ring signature algorithm provides a high degree of anonymity, due to the anonymity of the ring signature depending on the scale and diversity of the members within the ring, if the information of the members within the ring is correlated, the anonymity of the user identity will decrease, thereby posing a risk of leakage of the privacy of both trading parties. Summary of the Invention

[0004] In order to solve the above technical problems, this application provides a method for preserving all-process data of electronic signatures based on blockchain to solve the existing problems.

[0005] A method for preserving all-process data of electronic signatures based on blockchain in this application adopts the following technical solutions:

[0006] An embodiment of this application provides a method for preserving all-process data of electronic signatures based on blockchain, and this method includes the following steps:

[0007] Collect all historical transaction behavior data of each user before each transaction in the blockchain, including transaction users, transaction times, and text data of signed documents;

[0008] Analyze the similarity degree of the text data of any two transactions in the historical transaction behavior data, and the time interval between any two transactions, and determine the theme difference degree of the signed documents of any two transactions;

[0009] Determine the transaction behavior anomaly coefficient of each user at each transaction by combining the repetition degree of the transaction users in the historical transaction behavior data with the theme difference degree of all any two transactions;

[0010] Construct a transaction graph for each user's transaction based on the historical transaction behavior data of each user, analyze the degree of distribution chaos and the distribution range of the number of transactions between trading users in the transaction graph, and combine the transaction behavior anomaly coefficient to determine the malicious user confidence level for each user's transaction;

[0011] Identify and remove malicious users in the transaction graph of each user based on the malicious user confidence level; Denote the remaining users other than the user himself in the transaction graph after removing malicious users for each user's transaction as the users to be signed;

[0012] For each transaction, based on the number of transactions between each user and its users to be signed, and combining the malicious user confidence levels of the users to be signed, obtain the selectable coefficients of the users to be signed for each user's transaction, screen the users to be signed, and use an encryption signature algorithm to perform electronic signatures on the screened users to be signed for each user's transaction.

[0013] In one embodiment, the determination of the theme difference degree includes:

[0014] Use a word segmentation tool to obtain the keywords and their word frequencies in the text data of the signed document. Based on the word frequencies as weights, and based on the keywords in each signed document, use the Simhash algorithm to obtain the Simhash signatures of each signed document;

[0015] Calculate the metric distance between the Simhash signatures of the signed documents for any two transactions. The theme difference degree is positively correlated with the metric distance and negatively correlated with the time interval.

[0016] In one embodiment, the theme difference degree is the ratio of the metric distance to the time interval.

[0017] In one embodiment, the determination process of the transaction behavior anomaly coefficient is as follows:

[0018] For the historical transaction behavior data of each user's transaction, calculate the mean value of the theme difference degrees of all any two transactions, divide all the trading users in the historical transaction behavior data into two parts in the transaction order, and calculate the similarity coefficient between the two parts;

[0019] The transaction behavior anomaly coefficient is positively correlated with the mean value and negatively correlated with the similarity coefficient.

[0020] In one embodiment, the nodes in the transaction graph are each user, and the connection line between nodes is the normalized value of the number of transactions between the corresponding users, denoted as the transaction weight.

[0021] In one embodiment, the determination of the malicious user confidence level includes:

[0022] Form a transaction set with the transaction weights between each user and all other trading users in the transaction graph of each user, and calculate the product of the degree of chaos of all data in the transaction set and the average level;

[0023] Calculate the sum of the product and a preset value greater than 0, and the malicious user confidence level is the ratio of the transaction behavior anomaly coefficient to the sum.

[0024] In one embodiment, identifying and removing malicious users in the transaction graph of each user based on the malicious user confidence level includes:

[0025] Determine the trading users with a malicious user confidence level greater than a preset threshold among the trading users in the transaction graph of each user as malicious users, and remove them from the transaction graph of each user.

[0026] In one embodiment, the determination of the selectable coefficient includes:

[0027] Use the Kruskal algorithm to generate the minimum spanning tree of the transaction graph for each user's each transaction according to the principle of the smallest weight, and calculate the transaction separation degree between each user and the users to be signed in its minimum spanning tree for each user's each transaction based on the transaction weights;

[0028] Calculate the cumulative sum of the malicious user confidence levels of the users to be signed in the minimum spanning tree and a preset constant greater than 0, and the selectable coefficient of each user to be signed is the ratio of the transaction separation degree to the cumulative sum.

[0029] In one embodiment, the transaction separation degree is the cumulative sum of the reciprocals of all transaction weights between each user and the users to be signed in its minimum spanning tree for each user's each transaction.

[0030] In one embodiment, screening the users to be signed includes:

[0031] Arrange the selectable coefficients of all the users to be signed for each user's each transaction in descending order, and take the first preset number of users to be signed as the screened users to be signed.

[0032] This application has at least the following beneficial effects:

[0033] This application collects all historical transaction behavior data of each user before each transaction in the blockchain, including transaction users, transaction times, and text data of signed documents; analyzes the similarity degree of the text data of any two transactions in the historical transaction behavior data, and the time interval between the any two transactions, to determine the theme difference degree of the signed documents of the any two transactions; the theme difference degree reflects the difference degree of the text data of the signed documents in the historical transaction behavior data, thereby reflecting the possibility that the user belongs to an abnormal transaction; through the repetition degree of the transaction users in the historical transaction behavior data, combined with the theme difference degree of all any two transactions, determines the transaction behavior anomaly coefficient of each user at each transaction; the transaction behavior anomaly coefficient reflects the stability degree of the transaction users in the historical transaction behavior data of the user, reflects the credibility of the user's transaction behavior, and improves the accuracy of malicious user identification; constructs a transaction graph of each user at each transaction based on the historical transaction behavior data of each user, analyzes the degree of distribution chaos and the distribution range of the number of transactions between transaction users in the transaction graph, and combines the transaction behavior anomaly coefficient to determine the malicious user confidence level of each user at each transaction; the malicious user confidence level reflects the possibility that the user belongs to a malicious user, that is, an accurate assessment of the user's frequent abnormal transactions, which is beneficial to screening and obtaining safe and trustworthy users; identifies and eliminates malicious users in the transaction graph of each user based on the malicious user confidence level; records the remaining users other than the user himself in the transaction graph after eliminating malicious users for each user at each transaction as the users to be signed; for each transaction, based on the number of transactions between each user and its users to be signed, combined with the malicious user confidence level of the users to be signed, obtains the selectable coefficient of the users to be signed for each user at each transaction. The selectable coefficient considers the association relationship between the user to be signed and the signer, evaluates the possibility that the user to be signed becomes a member within the signer ring, and improves the security of the signer's electronic signature; and screens the users to be signed, and uses an encrypted signature algorithm to perform electronic signature on the screened users to be signed for each user at each transaction, so as to better ensure the anonymity of the signer's identity, and further protect the privacy and security of both parties to the transaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only 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.

[0035] Figure 1 It is a flowchart of the steps of a method for storing all-process data of electronic signature based on blockchain provided by the present application;

[0036] Figure 2 Schematic diagram of the minimum spanning tree for signer A;

[0037] Figure 3 Flowchart for determining selectable coefficients. Specific implementation manners

[0038] In order to further elaborate on the technical means and effects adopted by this application to achieve the predetermined invention purpose, the following combines the accompanying drawings and preferred embodiments to detail the specific implementation manners, structures, features and effects of a blockchain-based electronic signature whole-process data storage method proposed according to this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs.

[0040] The following specifically describes the specific solution of a blockchain-based electronic signature whole-process data storage method provided by this application with reference to the accompanying drawings.

[0041] A blockchain-based electronic signature whole-process data storage method provided by an embodiment of this application. Specifically, the following blockchain-based electronic signature whole-process data storage method is provided. Please refer to Figure 1 , and the method includes the following steps:

[0042] Step S001: Collect all historical transaction behavior data of each user before each transaction in the blockchain, including transaction users, transaction times, and text data of signed documents, and perform preprocessing.

[0043] The electronic signature data stored on the blockchain needs to consider the issue of user identity privacy. At present, due to the overly open and transparent nature of the blockchain, the user's identity privacy is maliciously obtained, bringing a poor experience to users. Therefore, this application aims to improve the blockchain-based electronic signature data storage method.

[0044] First, taking the nth transaction of each user in the blockchain as an example, collect the historical transaction behavior data of the previous n - 1 times before the nth transaction behavior of each user in the blockchain. Among them, the historical transaction behavior data includes the trading user, trading time, and text data of the signed document. Secondly, for the convenience of analysis, the trading users are processed using label encoding. In addition, in order to prevent incomplete information and data duplication in the collected data, the historical transaction behavior data of all users in the obtained blockchain is cleaned, and the data with incomplete information and duplicates is deleted. In this embodiment, the text data of the signed document is segmented using the Harbin Institute of Technology LTP tool, and keywords and word frequencies of the text data of the signed document are obtained based on the TF-IDF algorithm. Among them, label encoding, the Harbin Institute of Technology LTP tool, and the TF-IDF algorithm are all well-known technologies, and this embodiment will not elaborate on them in detail here.

[0045] Step S002, analyze the similarity degree of the text data of any two transactions in the historical transaction behavior data, and the time interval between the any two transactions, and determine the theme difference degree of the signed documents of the any two transactions.

[0046] One of the core features of the blockchain is transparent traceability. All transactions on the blockchain must be public to all nodes, and any node can verify the validity of the data. However, this feature of the blockchain also brings a problem, that is, excessive public transparency may lead to the leakage of users' privacy information. With the rapid popularization and promotion of blockchain applications, the risk of privacy leakage is becoming more and more serious. Therefore, this embodiment improves the insufficient privacy protection of the ring signature algorithm, and constructs a selectable coefficient of the user to be signed by analyzing the transaction behavior of users on the blockchain and the correlation between users, so as to improve the whole-process data evidence preservation of electronic signatures based on the blockchain.

[0047] First, considering that not all users on the blockchain are secure and trustworthy users, and there are many malicious users on the blockchain, this embodiment records any user who needs to sign as signer A. When signer A uses the ring signature algorithm for the When there is a malicious user among the ring members selected during the secondary signature, it may pose a threat to the security of transactions through forgery of signatures and other malicious attack behaviors. Therefore, it is necessary to eliminate malicious users in the blockchain. Since the business activities of secure and trustworthy users in the blockchain often have continuity and stability, that is, each user has a main business area, and the main business area of the user will not change significantly in a short period of time. This makes the content of the electronic signature files signed by secure and trustworthy users relatively stable in a short period of time, without significant differences, and the possibility of signing a large number of files in a short period of time is relatively low. For malicious users, they usually conduct transactions with users by quickly generating a large number of forged electronic signatures and the content of the signed files. As a result, the similarity of the content between two adjacent signed files of malicious users is relatively low.

[0048] Based on the above analysis, in this embodiment, user B in the blockchain is taken as an example for analysis. First, the keywords and word frequencies of the signed files for each transaction in the historical transaction behavior data of user B are obtained. The keywords of the signed files at each transaction are used as inputs, and the word frequencies corresponding to the keywords are used as weights. The Simhash signature of the signed files for each transaction is obtained by using the Simhash algorithm. Among them, the Simhash algorithm is a well-known existing technology, and the specific process will not be elaborated here. Calculate the theme difference degree between the signed files of any two transactions in the historical transaction behavior data of user B. The specific calculation method is as follows:

[0049] ; In the formula, is the theme difference degree between the signed files of the i-th and j-th transactions in the historical transaction behavior data of user B when conducting the n-th transaction, is the metric distance of the Simhash signatures of the signed files of the i-th and j-th transactions in the historical transaction behavior data of user B when conducting the n-th transaction, is the time interval between the i-th and j-th transactions in the historical transaction behavior data of user B when conducting the n-th transaction.

[0050] It should be noted that in this embodiment, the calculation method of the metric distance uses the Hamming distance. The calculation of the Hamming distance is a well-known existing technology, and implementers can choose other existing feasible calculation methods of the metric distance by themselves. This embodiment does not limit it here.

[0051] It should be understood that the smaller the metric distance, the higher the similarity degree between the two signed files, the smaller the difference degree, that is, the smaller the theme difference degree. If the time interval between two signed files of user B is closer, and the metric distance is larger, it indicates that the theme content difference of the files signed by user B is larger, and the possibility that user B is a malicious user is greater.

[0052] Step S003: Determine the transaction behavior anomaly coefficient for each user's each transaction by combining the repetition degree of the trading users in the historical transaction behavior data and the subject difference degree between any two transactions.

[0053] Secondly, in the blockchain, the trading users of secure and trustworthy users are relatively stable, and users conduct long-term transactions with each other. Malicious users usually frequently change trading users to avoid risks. Therefore, in this embodiment, the trading users in the historical transaction behavior data of user B when conducting the nth transaction are equally divided into two parts according to the transaction order. Among them, when the number of trading users in the historical transaction behavior data is odd, the trading user generated by the middle transaction behavior does not participate in the division. This embodiment uses the Jaccard algorithm to calculate the similarity coefficient between the two parts of trading users divided in the historical transaction behavior data of user B when conducting the nth transaction, and records it as the trading user consistency coefficient of user B. Among them, the larger the trading user consistency coefficient, the more stable the trading users of user B, and the more likely user B is a secure and trustworthy user. Jaccard is a well-known existing technology, and the specific process will not be elaborated. Implementers can choose other existing feasible calculation methods of similarity coefficients according to the actual situation, and this embodiment does not limit it here.

[0054] Based on the above analysis, calculate the transaction behavior anomaly coefficient for each user's each transaction. The specific calculation method is as follows:

[0055] ; In the formula, is the transaction behavior anomaly coefficient when the user conducts the nth transaction, is the subject difference degree between the signed documents of the ith and jth transactions in the historical transaction behavior data of user B when conducting the nth transaction, N is the number of transaction times in the historical transaction behavior data of user B when conducting the nth transaction, is the similarity coefficient between the two parts of trading users divided in the historical transaction behavior data of user B when conducting the nth transaction, is a preset value greater than 0 to avoid the denominator being 0. In this embodiment, , and implementers can set it by themselves according to the actual situation, and this embodiment does not limit it here.

[0056] It should be understood that when the content difference between the th signed document and the th signed document in the historical transaction behavior data of user B in the blockchain when conducting the th transaction is relatively large, the larger the Hamming distance value between the th signed document and the th signed document calculated; at the same time, the th signed document and the The closer the signing times of the signed documents are to each other, the smaller the value of the calculated time interval; and then the greater the value of the subject difference degree between the th signed document and the th signed document; indicating that the subject content of the signed documents in the historical transaction behavior data of user B during the th transaction is more different. Secondly, since malicious users frequently change trading users, the value of the trading user consistency coefficient calculated for the user is small; and then if the value of the transaction behavior anomaly coefficient of user B in the th transaction in the blockchain is large; it means that the risk degree of user B's trading mode is higher and it is more likely to be a potential malicious user.

[0057] Step S004: Based on the historical transaction behavior data of each user, construct a transaction graph for each user's each transaction, analyze the degree of distribution chaos and the distribution range of the number of transactions between trading users in the transaction graph, and combine the transaction behavior anomaly coefficient to determine the malicious user confidence level for each user's each transaction.

[0058] In the blockchain, the trading patterns between users are intricate, and there may be direct or potential association relationships between any two users. In this environment, when generating an electronic signature using the ring signature algorithm, the selection of ring members becomes crucial. When the selected ring members have a strong association with the signer, attackers are very likely to infer the true identity of the signer through techniques such as clustering analysis, thus destroying the anonymity of the signature. Therefore, this embodiment needs to analyze the association relationships between users on the blockchain.

[0059] Specifically, taking user B as an example again, construct a transaction graph according to the trading users in the historical transaction behavior data of user B in the blockchain. Among them, the purpose of constructing the transaction graph is to better present the potential trading relationships between users in the blockchain, and then analyze the association degree between user B and signer A in the blockchain. Each user is used as each node in the transaction graph, the nodes corresponding to the two users with transaction behaviors are connected, and the historical transaction times between the two connected users are normalized and used as the connection weight between the two nodes, denoted as the transaction weight, that is, the greater the transaction weight between the two users, the more transaction times there are between the two users, and the closer the association relationship between the two users. Among them, in this embodiment, the Sigmoid function is used to normalize the transaction times, and implementers can choose other existing feasible normalization methods by themselves.

[0060] Secondly, when the transaction behavior anomaly coefficient of user B in the blockchain is larger, it indicates that the possibility of user B being a malicious user is greater, and the probability of user B conducting abnormal transactions with other users is higher. Moreover, since malicious users usually do not choose to conduct multiple transactions with the same user, the number of transactions between user B and other users in the transaction graph is relatively close. That is, the transaction weights between the transaction nodes corresponding to user B and other users in the transaction graph are usually small and the differences are small. Therefore, in this embodiment, the set composed of the transaction weights between user B and all other trading users in the transaction graph of user B is denoted as the transaction set. Based on the data distribution in the transaction set and in combination with the transaction behavior anomaly coefficient, the malicious user confidence level when user B conducts the nth transaction is calculated. The specific calculation method is as follows:

[0061] ; In the formula, is the malicious user confidence level when user B conducts the nth transaction, is the transaction behavior anomaly coefficient when the user conducts the nth transaction, is the average value of all data in the transaction set when user B conducts the nth transaction, is the degree of chaos of all data in the transaction set when user B conducts the nth transaction, is a preset value greater than 0 to avoid the denominator being 0. In this embodiment, and the implementer can set it according to the actual situation. This embodiment does not limit it here.

[0062] It should be noted that in this embodiment, the calculation method of the degree of chaos adopts information entropy, and the implementer can choose other existing feasible calculation methods by himself, such as the coefficient of variation, standard deviation, etc.

[0063] It should be understood that when the possibility of user B having abnormal transactions in the blockchain is relatively large, the value of the transaction behavior anomaly coefficient when user B conducts the th transaction calculated is larger. And when user B conducts abnormal transactions with multiple users in the blockchain, the transaction weights between user B and other user nodes in the constructed transaction graph are usually small and basically equal in size. Furthermore, the average value and information entropy of the transaction set between user B and other user nodes in the transaction graph composed of the historical transaction behavior data of user B calculated are smaller. That is, the difference between the number of transactions between user B and other users is smaller and the number of transactions is smaller, and the value of the malicious user confidence level when user B conducts the th transaction obtained is larger, indicating that the possibility of user B being a malicious user is greater.

[0064] Step S005: Identify and remove malicious users in the transaction graph of each user based on the malicious user confidence level; Denote the remaining users other than the user himself / herself in the transaction graph after removing malicious users for each user's each transaction as the users to be signed.

[0065] By using the same calculation method for the malicious user confidence level when user B conducts the nth transaction, the malicious user confidence level of each user in the blockchain for each transaction can be obtained. For signer A, normalize the malicious user confidence level of each trading user in the transaction graph when signer A conducts the nth transaction, and determine the trading users with the normalized malicious user confidence level greater than the preset threshold as malicious users, and remove them from the transaction graph when signer A conducts the nth transaction to obtain a transaction graph composed of secure and trustworthy users for signer A.

[0066] In this embodiment, the Sigmoid function is used to normalize the malicious user confidence level. Implementers can choose other existing feasible normalization methods by themselves. The preset threshold in this embodiment is set to 0.8, and implementers can set it according to the actual situation by themselves.

[0067] Denote the transaction graph after removing malicious users when signer A conducts the nth transaction as transaction graph L, and denote the remaining users other than signer A in transaction graph L as the users to be signed.

[0068] Step S006: For each transaction, based on the number of transactions between each user and its users to be signed, and in combination with the malicious user confidence level of each user to be signed, obtain the selectable coefficient of each user to be signed for each user's each transaction, screen the users to be signed, and use an encryption signature algorithm to perform an electronic signature on the screened users to be signed for each user's each transaction.

[0069] This embodiment uses the Kruskal algorithm to generate the minimum spanning tree of the transaction graph The minimum spanning tree is generated according to the principle of the smallest transaction weight. The schematic diagram of the minimum spanning tree of signer A is as Figure 2 shown, Figure 2 in which represents the users in the blockchain, A represents signer A, and represents the transaction weight between users. Among them, the Kruskal algorithm is a well-known existing technology, and this embodiment will not elaborate on it in detail here. First, according to the minimum spanning tree, calculate the transaction separation degree between signer A and the users to be signed in its minimum spanning tree when signer A conducts the nth transaction. The specific calculation method is as follows:

[0070] ; In the formula, is the transaction separation degree between signer A and user F to be signed in its minimum spanning tree when signer A conducts the nth transaction, It is the k-th transaction weight between signer A and the user F to be signed in the minimum spanning tree during the n-th transaction of signer A, and K is the number of transaction weights between signer A and the user F to be signed in the minimum spanning tree during the n-th transaction of signer A.

[0071] Furthermore, calculate the selectable coefficients of each user to be signed during the n-th transaction of signer A. The specific calculation method is as follows:

[0072] ; In the formula, is the selectable coefficient of user F to be signed during the n-th transaction of signer A, is the transaction separation degree between signer A and user F to be signed in the minimum spanning tree during the n-th transaction of signer A, is the malicious user confidence level of user F to be signed during the n-th transaction of signer A, is a preset constant greater than 0 to avoid the denominator being 0. In this embodiment, , which can be set by the implementer according to the actual situation and is not limited in this embodiment. The flowchart for determining the selectable coefficient is as shown in Figure 3 .

[0073] It should be understood that when user F in the blockchain belongs to a secure and trustworthy user, the value of the malicious user confidence level of user F calculated is smaller; in addition, the smaller the direct and indirect transaction times between user F and signer A, the smaller the value of the direct or indirect transaction weight between user F and signer A calculated, and further the larger the value of the transaction separation degree between user F and signer A calculated; the larger the value of the selectable coefficient of user F to be signed finally calculated; indicating that the higher the security level of user F to be signed as a member of the ring of signer A. The selectable coefficient takes into account the trustworthiness of the user to be signed and the transaction separation degree between the user to be signed and signer A. According to the selectable coefficient, users in the ring members of signer A can be selected more securely.

[0074] Finally, considering that too many members in the ring will greatly increase the computational cost of the ring signature generation and verification process, while too few ring members will reduce the anonymity of the electronic signature. Therefore, in this embodiment, the selectable coefficients of all users to be signed during the n-th transaction of signer A are sorted in descending order, and the first preset number of users to be signed are formed into the ring members of signer A, and the ring signature algorithm is used to generate the electronic signature of signer A. In this embodiment, the preset number is denoted as g, and in this embodiment , where is the number of all users to be signed during the n-th transaction of signer A, is the ceiling function, and g is the number of members in the ring when signer A generates an electronic signature using the ring signature algorithm. Among them, the ring signature algorithm is a well-known existing technology, and this embodiment will not be elaborated here.

[0075] Store the generated electronic signature of signer A on the blockchain to complete the data deposit of the electronic signature.

[0076] By adopting the same operation steps as those for generating the electronic signature during the nth transaction of signer A, the electronic signatures generated by each user during each transaction in the blockchain can be obtained.

[0077] It should be noted that: the above order of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above specific embodiments of this specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0078] Each embodiment in this specification is described in a progressive manner. For the same or similar parts between each embodiment, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments.

[0079] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; modifying the technical solutions recorded in the foregoing embodiments, or equivalently replacing some of the technical features, does not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of each embodiment of the present application, and should all be included in the protection scope of the present application.

Claims

1. A blockchain-based method for storing data in the entire process of electronic signatures, characterized in that: The method comprises the following steps: Collect all historical transaction behavior data of each user in the blockchain before each transaction, including the transaction user, transaction time, and text data of the signed document; Analyze the similarity of the text data of any two transactions in the historical transaction behavior data, as well as the time interval between the two transactions, to determine the subject difference of the signed documents of the two transactions; Determining the abnormal coefficient of transaction behavior of each user at each transaction, including: for the historical transaction behavior data of each user at each transaction, calculating the average of the subject difference between all two arbitrary transactions, dividing all transaction users in the historical transaction behavior data into two parts according to the transaction order, and calculating the similarity coefficient between the two parts; the abnormal coefficient of transaction behavior is positively correlated with the average, and negatively correlated with the similarity coefficient; Based on the historical transaction behavior data of each user, a transaction graph of each user at each transaction is constructed, where the nodes in the transaction graph are each user, and the lines between the nodes are the normalized values ​​of the number of transactions between the corresponding users, recorded as transaction weights; the transaction weights of each user's transaction graph and all other transaction users are combined into a transaction set, and the product of the degree of chaos of all data in the transaction set and the average level is calculated; the sum of the product and a preset value greater than 0 is calculated, and the malicious user confidence is the ratio of the transaction behavior abnormality coefficient to the sum; Based on the malicious user confidence, malicious users in the transaction graph of each user are identified and eliminated; the remaining users in the transaction graph of each user after the malicious users are eliminated are recorded as users to be signed; For each transaction, based on the number of transactions between each user and its users to be signed, combined with the malicious user confidence of each user to be signed, the selectability coefficient of each user to be signed for each transaction is obtained, and the users to be signed are screened. For each transaction of each user, the screened users to be signed are electronically signed using an encrypted signature algorithm.

2. The method for storing data of the entire process of electronic signature based on blockchain as claimed in claim 1 is characterized in that: The determination of the subject difference degree includes: Use word segmentation tools to obtain keywords and their word frequencies in the text data of the signed documents, use word frequencies as weights, and use Simhash algorithm based on keywords in each signed document to obtain Simhash signatures of each signed document; The metric distance of the Simhash signatures of the signed files during the two arbitrary transactions is calculated, and the subject difference is positively correlated with the metric distance and negatively correlated with the time interval.

3. A blockchain-based electronic signature full-process data storage method as claimed in claim 2, characterized in that: The topic difference is the ratio of the metric distance to the time interval.

4. The method for storing data of the entire process of electronic signature based on blockchain as claimed in claim 1 is characterized in that: The identifying and removing malicious users from the transaction graph of each user based on the malicious user confidence level includes: Among the transaction users in the transaction graph of each user, transaction users whose malicious user confidence is greater than a preset threshold are determined as malicious users and are removed from the transaction graph of each user.

5. The method for storing data of the entire process of electronic signature based on blockchain as claimed in claim 1 is characterized in that: The determination of the selectable coefficient includes: The Kruskal algorithm is used to generate a minimum spanning tree of the transaction graph of each user at each transaction according to the minimum weight principle. Based on the transaction weight between each user at each transaction and the user to be signed in the minimum spanning tree, the transaction separation degree between each user at each transaction and the user to be signed in the minimum spanning tree is calculated; The cumulative sum of the malicious user confidence of the user to be signed in the minimum spanning tree and a preset constant greater than 0 is calculated, and the selectable coefficient of each user to be signed is the ratio of the transaction separation degree to the cumulative sum.

6. A blockchain-based electronic signature full-process data storage method as claimed in claim 5, characterized in that: The transaction separation degree is the cumulative sum of the reciprocals of all transaction weights between each user and the users to be signed in the minimum spanning tree at each transaction.

7. The method for storing data of the entire process of electronic signature based on blockchain as claimed in claim 1 is characterized in that: The user to be signed is screened, including: Arrange the selectable coefficients of all the users to be signed in descending order during each transaction of each user, and take the first preset number of users to be signed as the screened users to be signed.

Images