Block chain supervision method and system based on NLP and rule fusion

Abstract

The invention relates to the technical field of block chains, and particularly discloses a block chain supervision method and system based on NLP and rule fusion, and the method comprises the following steps: S1, an initialization stage: introducing an intelligent detection node and a supervision node into a block chain, and enabling the intelligent detection node and the supervision node to jointly participate in the transaction endorsement of the block chain; s2, in a supervision stage, the intelligent detection node and the supervision node inspect transaction contents in the current block and the historical block to obtain an inspection result, and the inspection result is reported to the decision node; and S3, in a decision-making stage, a decision-making node synthesizes the examination results of the intelligent detection node and the supervision node to generate a supervision conclusion, and carries out corresponding processing on illegal blocks.

Description

Technical Field

[0001] The present invention relates to the field of blockchain technology, and more specifically, to a blockchain supervision method and system based on the fusion of NLP and rules. Background Art

[0002] As a decentralized and tamper-proof distributed ledger technology, blockchain technology has broad application prospects. However, with the rapid development of blockchain technology, regulatory issues have gradually emerged and become an important factor restricting the healthy development of blockchain technology. The decentralized nature of blockchain makes it difficult to implement supervision. Traditional regulatory means are often helpless when facing blockchain, and it is difficult to effectively regulate blockchain. At the same time, the anonymity of blockchain technology also increases the difficulty of supervision, giving criminals an opportunity to take advantage. The complexity and continuous updates and iterations of blockchain technology make it difficult for regulatory technical means to keep up. Existing regulatory technical means are relatively backward and cannot effectively monitor and analyze blockchain.

[0003] At present, NLP technology is very mature in text classification processing. Compared with sensitive word matching methods, the introduction of NLP technology can improve the accuracy and efficiency of on-chain content review and reduce the false alarm rate. However, in actual application, the following problems still exist:

[0004] 1. NLP algorithms require high-quality, structured data for effective training and reasoning, but the data on the blockchain may come from different sources, and their quality and format may not be uniform, resulting in the immaturity of using NLP technology for on-chain content review.

[0005] 2. In the named entity recognition task, the sensitive word library matching technology performs poorly when dealing with abbreviations, aliases, and some ambiguous entities. In the sentiment analysis task, it is necessary to combine contextual information to understand the emotional tone of the text. However, since the sensitive word library matching technology only relies on complete words, it cannot capture the emotional signals contained in phrases or words.

[0006] In view of this, the present invention provides a blockchain supervision method and system based on the fusion of NLP and rules to solve the above problems. Summary of the invention

[0007] In order to overcome the problems in the prior art, the present invention proposes a blockchain supervision method and system based on the fusion of NLP and rules to solve the problem that traditional secret sharing technology relies on authoritative nodes, resulting in the centralized generation and distribution of trapdoors.

[0008] The present invention provides a blockchain supervision method based on NLP and rule fusion, comprising the following steps:

[0009] Step S1: In the initialization phase, smart detection nodes and regulatory nodes are introduced into the blockchain, and the smart detection nodes and regulatory nodes jointly participate in blockchain transaction endorsement;

[0010] Step S2: During the supervision phase, the intelligent detection node and the supervision node review the transaction content in the current block and the historical block, obtain the inspection results, and report to the decision node;

[0011] Step S3: In the decision-making stage, the decision node generates a supervision conclusion based on the review results of the intelligent detection node and the supervision node, and handles the illegal blocks accordingly.

[0012] As a preferred technical solution of the present invention, the specific steps of step S1 are:

[0013] Constructing a blockchain network, the blockchain network including distributed block nodes, including ordinary nodes, smart detection nodes and regulatory nodes; compiling regulatory rules and compliance requirements into smart contracts and deploying them in the blockchain network;

[0014] After deploying the smart contract, set the endorsement policy to ensure that the transaction is endorsed by enough nodes before being added to the blockchain.

[0015] As a preferred technical solution of the present invention, the transaction data in the current block and the historical block are parsed to extract the key information of each transaction.

[0016] Intelligent detection nodes and supervision nodes analyze the content of each transaction item by item through the rule base and NLP model, authenticate the sender and receiver of the transaction, and confirm whether they are trusted entities or are on the system's blacklist.

[0017] The NLP model is used to detect sensitive keywords in transaction notes or additional information, and the thresholds for transaction frequency and amount are set using the rule base. Transactions exceeding these thresholds will be marked as suspicious.

[0018] As a preferred technical solution of the present invention, in step S2, the intelligent detection node reviews the transaction content, including:

[0019] Extract useful text information from the transaction content and obtain word vectors based on the combined probability of each field in the text information;

[0020] A preset sensitive word library and rich historical corpus data are used to generate a sensitive word vector library; the vocabulary in the sensitive word library is combined with the context information in the corpus, and the sensitive word vector is obtained by training through the first algorithm.

[0021] The word vector of the input text is accurately matched with the pre-established sensitive word vector library to obtain the sensitive word vector.

[0022] As a preferred technical solution of the present invention, semantic analysis of transaction content is performed based on NLP technology, including:

[0023] Processing the given transaction text information into pinyin to generate its pinyin information;

[0024] Use the preset sensitive word library to match the pinyin information and find the pinyin information that matches the sensitive words;

[0025] According to the preset rules, the successfully matched pinyin information is converted into text information to be analyzed, and then semantic analysis is performed on it according to the specified rules.

[0026] As a preferred technical solution of the present invention, in step S3, the decision node processes the review results reported by the intelligent detection node and the supervision node in step S2, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

[0027] As a preferred technical solution of the present invention, the processing of the review results reported by the intelligent detection node and the supervision node refers to processing the results reported by the intelligent detection node and the supervision node respectively, including:

[0028] The results reported by the smart detection nodes are classified into violation levels, with a weight of 70% as the basis for the final judgment of whether the transaction is compliant;

[0029] The results reported by the regulatory nodes are classified into violation levels, with a weight of 30% as the basis for the final judgment of whether the transaction is compliant;

[0030] Based on the review results of the two nodes, the decision node decides on the transaction processing plan.

[0031] In the second aspect, the present invention provides a blockchain supervision system based on NLP and rule fusion, based on the implementation of the first aspect, including an intelligent detection node module, a supervision node module and a decision node module

[0032] The intelligent detection node module adds the intelligent detection node supervision node to the blockchain. The intelligent detection node and the supervision node participate in the transaction endorsement together and conduct compliance checks on the transaction content. Among them, the intelligent detection node converts the transaction content into text information, performs semantic and sentiment analysis on the transaction content based on NLP technology to obtain the review results, and reports the results to the decision node;

[0033] Supervision node module: The supervision node reviews the collection of transaction content in combination with the illegal content rule library. For transactions that have been on the chain, the content of the current block and historical blocks are stored in the cache database, and compliance is reviewed in different time periods. The review results are reported to the decision node;

[0034] Decision node module: The decision node processes the reporting results of the intelligent detection node and the supervision node, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

[0035] In a third aspect, the present invention provides an electronic device, comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0036] The processor executes the first aspect by calling a computer program stored in the memory.

[0037] In a fourth aspect, the present invention provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the first aspect.

[0038] The specific advantages of the present invention are as follows:

[0039] The content supervision system of the blockchain of the present invention combines advanced NLP technology. It not only relies on the traditional sensitive word matching mechanism, but also analyzes the semantics and the combination probability of each field in the transaction information to use the word vector comparison method to find out the non-compliant information in the transaction content, which greatly improves the efficiency and accuracy of detecting non-compliant content, and greatly improves the misjudgment rate of transaction content. By combining with the rule library, it can perform content supervision in a more customized manner, effectively ensuring the data security in the blockchain. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flowchart of a method for reviewing transaction content by an intelligent detection node according to an embodiment of the present invention;

[0041] Figure 2 This is a flow chart of a blockchain supervision method based on NLP and rule integration according to an embodiment of the present invention;

[0042] Figure 3 This is a structural diagram of the blockchain supervision system based on the fusion of NLP and rules in an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Example 1

[0045] See also Figure 1-3 This embodiment provides a blockchain supervision method based on the fusion of NLP and rules, including the following steps:

[0046] Step S1: In the initialization phase, smart detection nodes and regulatory nodes are introduced into the blockchain, and the smart detection nodes and regulatory nodes jointly participate in blockchain transaction endorsement.

[0047] Specifically, the specific steps of step S1 are:

[0048] Constructing a blockchain network, the blockchain network including distributed block nodes, including ordinary nodes, smart detection nodes and regulatory nodes; compiling regulatory rules and compliance requirements into smart contracts and deploying them in the blockchain network;

[0049] After deploying the smart contract, set the endorsement policy to ensure that the transaction is endorsed by enough nodes before being added to the blockchain.

[0050] For example, you can set "AND('Org1 MSP.member','Org2MSP.member')" as the endorsement policy, which means that the transaction needs to be endorsed by members of both Org1 and Org2 to be verified. It also provides an extended function based on Go sh im, allowing chaincode developers to handle the endorsement policy based on the organization's MSP identifier, which can be completed by calling the SDK inside the smart contract.

[0051] In step S1, when a transaction starts, the introduced intelligent detection nodes and regulatory nodes will participate in the transaction endorsement together with other nodes in the blockchain. After the intelligent detection nodes and regulatory nodes complete their respective compliance analysis, they will form a review report or compliance assessment score and report the results to the decision-making node. Other ordinary nodes will perform basic transaction verification, including whether the account balance is sufficient, whether the transaction format is correct, etc. Ordinary nodes will also participate in the endorsement process, but are not responsible for in-depth compliance review. During the transaction endorsement process, the endorsement information of each node will include the node's verification results for the transaction. Ordinary nodes only verify whether the transaction meets the basic requirements of the blockchain protocol, while intelligent detection nodes and regulatory nodes will additionally record the detailed results of the compliance review and report the detailed results of the review to the decision-making node;

[0052] Step S2: In the group supervision stage, the intelligent detection node and the supervision node review the transaction content in the current block and the historical block, obtain the inspection results, and report to the decision node.

[0053] Specifically, in step S2, the analysis logic of the transaction content is:

[0054] Parse the transaction data in the current block and historical blocks to extract the key information of each transaction, including transaction ID, sender, receiver, amount and transaction notes. For each transaction, the intelligent detection node and the supervision node analyze its content item by item through the rule base and NLP model. Authenticate the sender and receiver of the transaction to confirm whether they are trusted entities or whether they are on the system's blacklist. Use the NLP model to detect sensitive keywords in transaction notes or additional information. Use the rule base to set thresholds for transaction frequency and amount. Transactions exceeding these thresholds will be marked as suspicious.

[0055] More specifically, in step S2, the transaction content is reviewed based on the intelligent detection node, including the following steps:

[0056] Extract useful text information from transaction content;

[0057] A high-quality sensitive word vector library is generated by using a preset sensitive word library and rich historical corpus data. The vocabulary in the sensitive word library is combined with the contextual information in the corpus, and trained through specific algorithms (such as Word2Vec, GloVe, etc.) to generate a word vector representation for each sensitive word.

[0058] Analyze the combined probability of each field in the text information to obtain the word vector;

[0059] A word vector comparison method is used to accurately match the word vector of the input text with the pre-established sensitive word vector library. In this process, we identify word vectors with the same or similar features as those in the sensitive word vector library and define them as the sensitive word vectors to be analyzed.

[0060] Based on NLP technology, semantic analysis of transaction content is performed to extract useful text information, including the following steps:

[0061] Processing the given transaction text information into pinyin to generate its pinyin information;

[0062] Use the preset sensitive word library to match the pinyin information, and find the part that matches the sensitive word, which is the pinyin information to be analyzed;

[0063] According to the preset rules, the successfully matched pinyin information is converted into text information to be analyzed, and then semantic analysis is performed on it according to the specified rules.

[0064] In step S2, the transaction content is reviewed based on the supervision node, including the following steps:

[0065] The supervision node reviews the transaction content. The supervision node combines the corresponding content rule library and reviews the transaction content according to the non-compliant content rule library, assesses the violation level, and reports the review results to the decision node.

[0066] The content rule library sets different upper limits for amounts based on transaction types and account types. Transactions exceeding the upper limit are marked as suspicious. An upper limit for transaction frequency within a unit of time is set. Frequent small transactions may be considered as risk diversification behavior. For transaction remarks or additional information, a list of sensitive keywords, such as "money laundering" and "tax evasion", is used to automatically detect whether these keywords are included through the NLP model.

[0067] Step S3: In the decision-making stage, the decision node generates a supervision conclusion based on the review results of the intelligent detection node and the supervision node, and handles the illegal blocks accordingly.

[0068] Specifically, in step S3, the decision node processes the review results reported by the intelligent detection node and the supervision node in step S2, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

[0069] Processing the review results reported by the intelligent detection node and the supervision node refers to processing the results reported by the intelligent detection node and the supervision node respectively, including:

[0070] The results reported by the smart detection nodes are classified into violation levels, with a weight of 70% as the basis for the final judgment of whether the transaction is compliant;

[0071] The results reported by the regulatory nodes are classified into violation levels, with a weight of 30% as the basis for the final judgment of whether the transaction is compliant;

[0072] Based on the review results of the two nodes, the decision node decides on the transaction processing plan.

[0073] Example 2: The part not described in detail in this example is as shown in Example 1. A specific example is provided for analyzing a blockchain supervision method based on the fusion of NLP and rules. In view of the complexity and diversified expression requirements of smart contract transaction information in the blockchain, refined processing steps are added to improve the accurate detection of sensitive information, including the following steps:

[0074] Step S101, extracting useful text information from the transaction content;

[0075] In this embodiment, the transaction content in the blockchain is based on the smart contract as the information carrier. A preprocessing module is introduced to analyze the unique structures such as contract call parameters and event logs of the blockchain data, structure the key information, and extract the transaction content in the smart contract as standardized text information.

[0076] As an optional embodiment, between step S101 and step S102, the following steps may also be included: performing similarity analysis on the text information, specifically, performing pinyin processing on the given transaction text information to generate its pinyin information; matching the pinyin information with a preset sensitive word library, and finding the part that matches the sensitive word, which is the pinyin information to be analyzed. Different from conventional pinyin matching, this embodiment introduces fuzzy matching of similar-sounding words. Through the fuzzy matching algorithm, the system can identify words with similar but not completely identical pinyin, and capture more potential sensitive expressions.

[0077] Among them, for the sensitive word vector library, a high-quality sensitive word vector library is generated using a preset sensitive word library and rich historical corpus data. Unlike conventional methods, the training of word vectors in this method includes contextual feature tags of the corpus. The semantic differences of specific sensitive words in different contexts such as financial transactions and resource allocation are recorded, which improves the scene adaptability of the detection results. The vocabulary in the sensitive word library is combined with the contextual information in the corpus, and trained through specific algorithms (such as Word2Vec, G loVe, etc.) to generate word vector representations for each sensitive word.

[0078] Step S102: perform semantic analysis on the transaction text information based on NLP technology and obtain analysis results.

[0079] In this embodiment, step S102 specifically includes: analyzing the combination probability of each field in the text information to obtain a word vector. When analyzing the field combination, this embodiment further introduces a mechanism for dynamic combination weight adjustment, so that the system can adjust the semantic weights of certain words in real time according to different field combinations. This feature enables the system to have more flexible recognition capabilities in specific transaction types; using a word vector comparison method, the word vector of the input text is accurately matched with a pre-established sensitive word vector library. In this embodiment, a contextual semantic comparison module is added. This module not only considers the similarity of the words themselves when performing word vector comparison, but also compares the word associations in the context. By comparing the lexical relationships in the context, the system can identify implicit sensitive content, even if its wording is more concealed.

[0080] Step S103, obtaining the review result.

[0081] In this embodiment, obtaining the review result means grading the reviewed transaction content. To make the decision node process the reviewed transaction, this embodiment adds a dynamic grading adjustment mechanism to enhance the flexibility of supervision. The system will dynamically adjust the grading based on the word vector similarity, strong context relevance, and the position of sensitive words in the sentence during the review process. If a sensitive word appears as a core word in the transaction information, the grading will be increased by one level to ensure that the system processes important sensitive information with high priority.

[0082] There are three levels of classification: minor, medium, and severe. For the three levels of review results, the decision results will be handled accordingly. Specifically, if the review result is minor, no action will be taken; if the review result is medium, partial blocking will be performed, such as improper remarks or sensitive words in the transaction content; if the review result is severe, the transaction will be rejected or all transaction content will be blocked.

[0083] Embodiment 3: This embodiment provides a blockchain supervision system based on NLP and rule fusion, including an intelligent detection node module, a supervision node module and a decision node module;

[0084] The intelligent detection node module adds the intelligent detection node supervision node to the blockchain. The intelligent detection node and the supervision node participate in the transaction endorsement together and conduct compliance checks on the transaction content. Among them, the intelligent detection node converts the transaction content into text information, performs semantic and sentiment analysis on the transaction content based on NLP technology to obtain the review results, and reports the results to the decision node;

[0085] Supervision node module: The supervision node reviews the collection of transaction content in combination with the illegal content rule library. For transactions that have been on the chain, the content of the current block and historical blocks are stored in the cache database, and compliance is reviewed in different time periods. The review results are reported to the decision node;

[0086] Decision node module: The decision node processes the reporting results of the intelligent detection node and the supervision node, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

[0087] It should be noted that the blockchain content supervision solution based on the integration of NLP and rules combines NLP technology with the knowledge rule base and integrates it into the supervision of blockchain, making blockchain supervision more precise and intelligent, and fundamentally reducing the adverse effects caused by non-compliant information on the chain.

[0088] Embodiment 3: For the parts not described in detail in this embodiment, please refer to the description of Embodiment 1. This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0089] The processor executes the above-mentioned blockchain supervision method based on the fusion of NLP and rules by calling the computer program stored in the memory.

[0090] Embodiment 4: This embodiment provides a computer-readable storage medium storing instructions. When the instructions are executed on a computer, the computer executes the above-mentioned blockchain supervision method based on the fusion of NLP and rules.

[0091] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0092] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A blockchain supervision method based on the fusion of NLP and rules, characterized in that: The following steps are involved: Step S1: In the initialization phase, smart detection nodes and regulatory nodes are introduced into the blockchain, and the smart detection nodes and regulatory nodes jointly participate in blockchain transaction endorsement; Step S2: During the supervision phase, the intelligent detection node and the supervision node review the transaction content in the current block and the historical block, obtain the inspection results, and report to the decision node; Step S3: In the decision-making stage, the decision node generates a supervision conclusion based on the review results of the intelligent detection node and the supervision node, and handles the illegal blocks accordingly.

2. According to claim 1, a blockchain supervision method based on NLP and rule fusion is characterized in that: The specific steps of step S1 are: Constructing a blockchain network, the blockchain network including distributed block nodes, including ordinary nodes, smart detection nodes and regulatory nodes; compiling regulatory rules and compliance requirements into smart contracts and deploying them in the blockchain network; After deploying the smart contract, set the endorsement policy to ensure that the transaction is endorsed by enough nodes before being added to the blockchain.

3. According to claim 2, a blockchain supervision method based on NLP and rule fusion is characterized in that: In step S2, the analysis logic of the transaction content is: Parse the transaction data in the current block and historical blocks to extract the key information of each transaction. Intelligent detection nodes and supervision nodes analyze the content of each transaction item by item through the rule base and NLP model, authenticate the sender and receiver of the transaction, and confirm whether they are trusted entities or are on the system's blacklist. The NLP model is used to detect sensitive keywords in transaction notes or additional information, and the thresholds for transaction frequency and amount are set using the rule base. Transactions exceeding these thresholds will be marked as suspicious.

4. According to claim 3, a blockchain supervision method based on NLP and rule fusion is characterized in that: In step S2, the intelligent detection node reviews the transaction content, including: Extract useful text information from the transaction content and obtain word vectors based on the combined probability of each field in the text information; A preset sensitive word library and rich historical corpus data are used to generate a sensitive word vector library; the vocabulary in the sensitive word library is combined with the context information in the corpus, and the sensitive word vector is obtained by training through the first algorithm. The word vector of the input text is accurately matched with the pre-established sensitive word vector library to obtain the sensitive word vector.

5. According to claim 4, a blockchain supervision method based on NLP and rule fusion is characterized in that: Perform semantic analysis of transaction content based on NLP technology, including: Processing the given transaction text information into pinyin to generate its pinyin information; Use the preset sensitive word library to match the pinyin information and find the pinyin information that matches the sensitive words; According to the preset rules, the successfully matched pinyin information is converted into text information to be analyzed, and then semantic analysis is performed on it according to the specified rules.

6. According to claim 5, a blockchain supervision method based on NLP and rule fusion is characterized in that: In step S3, the decision node processes the review results reported by the intelligent detection node and the supervision node in step S2, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

7. The blockchain supervision method based on NLP and rule fusion according to claim 6 is characterized in that: Processing the review results reported by the intelligent detection node and the supervision node refers to processing the results reported by the intelligent detection node and the supervision node respectively, including: The results reported by the smart detection nodes are classified into violation levels, with a weight of 70% as the basis for the final judgment of whether the transaction is compliant; The results reported by the regulatory nodes are classified into violation levels, with a weight of 30% as the basis for the final judgment of whether the transaction is compliant; Based on the review results of the two nodes, the decision node decides on the transaction processing plan.

8. The blockchain supervision method based on NLP and rule fusion according to claim 1 is characterized in that: The following steps are involved: Step S101, extracting useful text information from the transaction content, wherein the transaction content uses the smart contract as the information carrier; Step S102: semantic analysis is performed on the transaction text information based on NLP technology to obtain analysis results. When analyzing field combinations, a dynamic combination weight adjustment mechanism is introduced to adjust the semantic weights of certain words in real time according to different field combinations. Step S103, obtaining the review result, which includes light, medium and severe, and making corresponding processing according to the review result.

9. The blockchain supervision method based on NLP and rule fusion according to claim 8 is characterized in that: Between step S101 and step S102, the following steps are also included: Perform similarity analysis on text information and perform phonetic processing on given transaction text information to generate its phonetic information; Use the preset sensitive word library to match the pinyin information, find the part that matches the sensitive word, and introduce fuzzy matching of similar-sounding words; Through the fuzzy matching algorithm, the system can identify words with similar but not identical pinyin, capture more potential sensitive expressions, and generate a high-quality sensitive word vector library using the preset sensitive word library and rich historical corpus data for sensitive word vector libraries; The vocabulary in the sensitive word library is combined with the contextual information in the corpus and trained through an algorithm to generate a word vector representation for each sensitive word.

10. A blockchain supervision system based on NLP and rule fusion, based on the implementation of a blockchain supervision method based on NLP and rule fusion according to any one of claims 1 to 9, characterized in that: Including intelligent detection node module, supervision node module and decision node module The intelligent detection node module adds the intelligent detection node supervision node to the blockchain. The intelligent detection node and the supervision node participate in the transaction endorsement together and conduct compliance checks on the transaction content. Among them, the intelligent detection node converts the transaction content into text information, performs semantic and sentiment analysis on the transaction content based on NLP technology to obtain the review results, and reports the results to the decision node; Supervision node module: The supervision node reviews the collection of transaction content in combination with the illegal content rule library. For transactions that have been on the chain, the content of the current block and historical blocks are stored in the cache database, and compliance is reviewed in different time periods. The review results are reported to the decision node; Decision node module: The decision node processes the reporting results of the intelligent detection node and the supervision node, refuses to upload transactions with non-compliant content to the chain and stores them in the database for identification, and blocks or refuses to query the transaction content that has been uploaded to the chain.

Images