Blockchain-based industrial transaction digitalization risk supervision and early warning system

Abstract

The application discloses a blockchain-based industrial transaction digitization risk supervision and early warning system, and belongs to the technical field of financial transactions, and comprises: a blockchain storage module, which is used for storing industrial transaction whole-process data based on blockchain technology; a risk supervision identification module, which is used for constructing an industrial transaction digitization risk identification model to analyze and identify the industrial transaction whole-process data, and determining an industrial transaction digitization risk identification result; and a risk early warning management module, which is used for performing timely risk early warning according to the industrial transaction digitization risk identification result, and forming a digital closed-loop management of the industrial transaction. The application solves the problem that the existing industrial transaction cannot realize real-time monitoring, risk identification and early warning disposal of transaction data based on the blockchain, and leads to poor industrial transaction digitization management effect. The application can realize real-time monitoring, risk identification and early warning disposal of transaction data based on the blockchain, and can improve the industrial transaction digitization management effect.

Description

Technical Field

[0001] This invention relates to the field of financial transaction technology, specifically to a blockchain-based digital risk monitoring and early warning system for industrial transactions. Background Technology

[0002] Industrial transactions refer to a series of trading activities centered around the industrial chain, including raw material procurement, production and manufacturing, and sales and distribution. They encompass not only the exchange of goods and services between enterprises but also the flow of resources such as technology, information, and capital. With the rapid development of digital technology, industrial transactions are gradually transforming towards online and intelligent models.

[0003] Among these features, integrating the supply chain through a digital platform breaks down time and geographical limitations, enabling efficient transactions; utilizing big data to analyze market demand, optimize production plans and supply chain management, and improve decision-making efficiency; using blockchain technology to achieve transparency and tamper-proof transaction data, enhancing trust and security; and tracking the location and status of goods in real time, improving logistics efficiency and reducing costs.

[0004] However, existing industrial transactions cannot achieve real-time monitoring of transaction data, risk identification, and early warning and handling based on blockchain, resulting in poor digital management of industrial transactions. Summary of the Invention

[0005] The purpose of this invention is to provide a blockchain-based digital risk monitoring and early warning system for industrial transactions. This system can achieve real-time monitoring, risk identification, and early warning of transaction data based on blockchain, thereby improving the effectiveness of digital management of industrial transactions and solving the problems mentioned in the background.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A blockchain-based digital risk monitoring and early warning system for industrial transactions includes:

[0008] The blockchain storage module is used to store data of the entire industrial transaction process based on blockchain technology, and combines smart contracts to realize the automated execution of regulatory rules, ensuring that the data of the entire industrial transaction process is tamper-proof.

[0009] The risk supervision and identification module is used to build a digital risk identification model for industrial transactions, analyze and identify data throughout the entire industrial transaction process, and determine the results of digital risk identification for industrial transactions.

[0010] The risk warning and management module is used to provide timely risk warnings based on the results of digital risk identification in industrial transactions, and to form a digital closed-loop management of industrial transactions.

[0011] Preferably, based on blockchain technology, the entire process of industry transactions is stored, and the following operations are performed:

[0012] Based on smart contracts, data records are generated from industry transaction data, and the data records are packaged into blocks, with timestamps and the hash value of the previous block added.

[0013] Based on consensus algorithms, nodes in the blockchain network verify transactions in new blocks. Once verified, the new block is broadcast to the entire blockchain network, enabling all nodes in the blockchain network to update their ledgers synchronously.

[0014] New blocks are added to the blockchain with timestamps to form an immutable data chain, thereby generating full-process data of industrial transactions. The data records of each transaction can be queried through a blockchain explorer, which facilitates subsequent auditing and supervision.

[0015] Preferably, based on a consensus algorithm, nodes in the blockchain network verify transactions in a new block and perform the following operations:

[0016] The block header data of a new block includes the hash value of the previous block, the timestamp of the current block, the Merkle root of the transaction data, the random number, and the network's preset difficulty target;

[0017] The block header data is hashed based on a hash algorithm. The random number is continuously adjusted and the hash value is recalculated until a hash value that meets the difficulty requirement is found.

[0018] Once a hash value that meets the difficulty requirement is found, the new block is broadcast to the entire network. When other nodes receive the new block, they verify its hash value, the integrity of the transaction data, and the timestamp information to ensure the validity of the new block, thereby completing the transaction verification of the new block.

[0019] Preferred options also include:

[0020] Real-time monitoring of difficulty-related data is performed and compared with the corresponding preset safety threshold range to obtain safety comparison results.

[0021] Real-time monitoring of difficulty-related data is performed and compared with the corresponding preset safety threshold range to obtain safety comparison results.

[0022] Based on the security comparison results, if there is difficulty-related data that exceeds the preset security threshold range, then corresponding difficulty adjustment measures are triggered to adjust the difficulty target of the new block and generate a difficulty adjustment record.

[0023] If there is no difficulty-related data exceeding the preset security threshold, the current new block will be divided according to the preset division ratio to obtain micro-blocks;

[0024] If the generation time difference of adjacent preset micro-blocks is less than the set low time difference interval, the generation difficulty of the signal is increased.

[0025] If the generation time difference of adjacent preset micro-blocks is greater than the set high time difference interval, the generation difficulty of the signal is reduced.

[0026] Compare and analyze the frequency of occurrences of signals indicating increased difficulty in the current new block with the frequency of occurrences of signals indicating decreased difficulty, and determine the difference in signal frequency and the initial adjustment direction;

[0027] If the difference in the number of signals is not greater than the set threshold for the difference in the number of signals, then the difficulty target for the current new block will not be adjusted.

[0028] If the difference in the number of signal occurrences is greater than a set threshold, the current new block will be marked as a potentially abnormal block.

[0029] The historical difficulty-related data and historical difficulty adjustment records of the potentially abnormal blocks within a preset time period are input into the difficulty-related data prediction model for prediction, thereby obtaining the potentially difficult-related data.

[0030] If no data related to the potential difficulty exceeds the corresponding preset safety threshold, the difficulty target of the potentially abnormal block is adjusted based on the difference in the number of signals and the initial adjustment direction.

[0031] If there are any difficulty-related data that may exceed the corresponding preset safety threshold range, then the corresponding difficulty adjustment measures will be matched and regarded as reference adjustment measures;

[0032] Based on the aforementioned reference adjustment measures, the predicted adjustment direction for the difficulty target of the potentially anomalous blocks is determined;

[0033] When the initial adjustment direction and the predicted adjustment direction are consistent, the difficulty target of the potentially abnormal block is adjusted based on the reference adjustment measures;

[0034] When the initial adjustment direction and the predicted adjustment direction are inconsistent, the reference adjustment measures are adjusted based on the difference in the number of signal occurrences, and then the difficulty target of the possible abnormal block is adjusted.

[0035] Preferably, a digital risk identification model for industrial transactions is constructed, and the following operations are performed:

[0036] Collect historical data on industry transactions and divide the collected historical data into training set and test set.

[0037] The deep learning model is trained using a training set, enabling it to autonomously learn the risk identification behavior of digital industrial transactions from the training set and effectively identify abnormal behaviors or potential risks in digital industrial transactions, thus establishing a deep learning-based risk identification model for digital industrial transactions.

[0038] The test set was used to perform performance testing on the deep learning-based risk identification model for digital industrial transactions, to evaluate whether the deep learning-based risk identification model for digital industrial transactions can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions, and to determine the model test evaluation results.

[0039] Based on the model testing and evaluation results, the parameters of the deep learning-based industrial transaction digital risk identification model are adjusted and optimized to determine the optimal industrial transaction digital risk identification model that can effectively identify abnormal behaviors or potential risks in industrial transaction digitalization.

[0040] Preferably, the performance of the deep learning-based digital risk identification model for industrial transactions is tested using a test set, and the following operations are performed:

[0041] The test set is input into the deep learning-based digital risk identification model for industrial transactions. The risk identification performance of the deep learning-based digital risk identification model for industrial transactions is tested based on the test set, and the real-time test data of the model is determined.

[0042] The model test real-time data is compared with the preset model test threshold data to analyze the matching between the model test real-time data and the model test threshold data, and to determine whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0043] When the real-time data of the model test matches the threshold data of the model test, the deep learning-based risk warning model for digital industrial transactions can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions.

[0044] When the real-time data of the model test does not match the threshold data of the model test, the risk warning model for digital industrial transactions based on deep learning cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in the digital industrial transactions.

[0045] Preferably, determining whether a deep learning-based risk identification model for digital industrial transactions can effectively identify abnormal behaviors or potential risks in digital industrial transactions also includes:

[0046] When there is no real-time model test data exceeding the preset model test threshold, it is determined that the current deep learning-based industrial transaction digitization risk warning model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0047] When there is real-time model test data that exceeds the preset model test threshold, the corresponding real-time model test data will be marked as matching analysis data.

[0048] The test matching coefficient is determined by combining the data difference between the matching analysis data and the corresponding preset model test threshold data with the corresponding performance feedback weight.

[0049] When the test matching coefficient exceeds the preset matching threshold, it is determined that the current deep learning-based industrial transaction digitization risk warning model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0050] Otherwise, it can be determined that the current deep learning-based risk warning model for digital industrial transactions cannot achieve the expected effect of effectively identifying abnormal behaviors or potential risks in the digitalization of industrial transactions.

[0051] Preferably, based on the model testing and evaluation results, the parameters of the deep learning-based industrial transaction digital risk identification model are adjusted and optimized, and the following operations are performed:

[0052] When the deep learning-based risk identification model for digital industrial transactions fails to achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions, the parameters of the deep learning-based risk identification model for digital industrial transactions will be adjusted and optimized.

[0053] The performance of the adjusted and optimized industrial transaction digital risk identification model will be tested again to re-evaluate whether the deep learning-based industrial transaction digital risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitalization. This process will continue until the deep learning-based industrial transaction digital risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitalization, thereby determining the optimal industrial transaction digital risk identification model for effectively identifying abnormal behavior or potential risks in industrial transaction digitalization.

[0054] Preferably, the data from the entire industrial transaction process is analyzed and identified to determine the digital risk identification results of industrial transactions, and the following operations are performed:

[0055] The entire process of industrial transactions is input into the industrial transaction digital risk identification model. The model analyzes the entire process of industrial transactions and automatically identifies abnormal behaviors or potential risks in the digitalization of industrial transactions, thereby determining the results of the digital risk identification of industrial transactions.

[0056] Preferably, based on the results of digital risk identification in industrial transactions, timely risk warnings are issued, and the following operations are performed:

[0057] Based on the results of risk identification in the digitalization of industrial transactions, timely risk warnings are issued and the risk warning information is pushed to the management in real time. The management can then manage abnormal behaviors or potential risks in the digitalization of industrial transactions in a timely manner, track the management situation throughout the entire process, and make real-time dynamic adjustments based on the tracking feedback, thus forming a closed-loop management of digitalization of industrial transactions.

[0058] Compared with the prior art, the beneficial effects of the present invention are:

[0059] This invention stores data from the entire industrial transaction process based on blockchain technology and combines it with smart contracts to automate the execution of regulatory rules, ensuring that the data from the entire industrial transaction process is tamper-proof. By constructing an industrial transaction digital risk identification model, it analyzes the data from the entire industrial transaction process and automatically identifies abnormal behaviors or potential risks in the digitalization of industrial transactions, thereby determining the results of the digital risk identification of industrial transactions. Based on the results of the digital risk identification of industrial transactions, timely risk warnings are issued, and a digital closed-loop management of industrial transactions is formed. It can realize real-time monitoring of transaction data, risk identification, and early warning and handling based on blockchain, which can improve the effectiveness of digital management of industrial transactions. Attached Figure Description

[0060] Figure 1 This is a block diagram of the digital risk monitoring and early warning system for industrial transactions of the present invention;

[0061] Figure 2 This is a flowchart of the digital risk monitoring and early warning system for industrial transactions according to the present invention. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] To address the issue of poor digital management of industrial transactions due to the inability of existing industrial transactions to achieve real-time monitoring, risk identification, and early warning based on blockchain technology, please refer to [link / reference needed]. Figures 1-2 This embodiment provides the following technical solution:

[0064] A blockchain-based digital risk monitoring and early warning system for industrial transactions includes:

[0065] The blockchain storage module is used to store data from the entire industrial transaction process based on blockchain technology, and combines smart contracts to achieve automated execution of regulatory rules, ensuring that the data from the entire industrial transaction process is tamper-proof.

[0066] In this embodiment, blockchain technology is used to store data from the entire industry transaction process, and the following operations are performed:

[0067] Based on smart contracts, data records are generated from industry transaction data, and the data records are packaged into blocks, with timestamps and the hash value of the previous block added.

[0068] Based on consensus algorithms, nodes in the blockchain network verify transactions in new blocks. Once verified, the new block is broadcast to the entire blockchain network, enabling all nodes in the blockchain network to update their ledgers synchronously.

[0069] New blocks are added to the blockchain with timestamps to form an immutable data chain, thereby generating full-process data of industrial transactions. The data records of each transaction can be queried through a blockchain explorer, which facilitates subsequent auditing and supervision.

[0070] It should be noted that blockchain, through distributed ledger technology, stores data on multiple nodes, avoiding reliance on a single centralized institution, thereby improving data security and reliability. Once data is recorded on the blockchain, it cannot be modified unless more than 51% of the network's computing power is controlled. This characteristic ensures the authenticity and integrity of the data. Furthermore, all transaction records on the blockchain are open and transparent to all participants. Any node can query and verify the data, but cannot arbitrarily tamper with it. Each transaction is linked to the previous blockchain through timestamps and hash values, forming an immutable chain that facilitates tracing the historical record of each transaction.

[0071] In this embodiment, based on the consensus algorithm, nodes in the blockchain network verify transactions in a new block and perform the following operations:

[0072] The block header data of a new block includes the hash value of the previous block, the timestamp of the current block, the Merkle root of the transaction data, the random number, and the network's preset difficulty target;

[0073] The block header data is hashed based on a hash algorithm. The random number is continuously adjusted and the hash value is recalculated until a hash value that meets the difficulty requirement is found.

[0074] Once a hash value that meets the difficulty requirement is found, the new block is broadcast to the entire network. When other nodes receive the new block, they verify its hash value, the integrity of the transaction data, and the timestamp information to ensure the validity of the new block, thereby completing the transaction verification of the new block.

[0075] It should be noted that a consensus algorithm is a mechanism that ensures all nodes in a blockchain network reach a consensus on the transaction order and ledger state. Its main functions include:

[0076] Ensure transaction consistency: Ensure that all nodes have a consistent view of transaction records and ledger status to avoid system crashes due to data conflicts;

[0077] Preventing fraud: Through a distributed consensus mechanism, it prevents malicious nodes from tampering with data or launching double-spending attacks.

[0078] The risk supervision and identification module is used to build a digital risk identification model for industrial transactions, analyze and identify data throughout the entire industrial transaction process, and determine the results of digital risk identification for industrial transactions.

[0079] In this embodiment, it also includes:

[0080] Real-time monitoring of difficulty-related data is performed and compared with the corresponding preset safety threshold range to obtain safety comparison results.

[0081] Based on the security comparison results, if there is difficulty-related data that exceeds the preset security threshold range, then corresponding difficulty adjustment measures are triggered to adjust the difficulty target of the new block and generate a difficulty adjustment record.

[0082] If there is no difficulty-related data exceeding the preset security threshold, the current new block will be divided according to the preset division ratio to obtain micro-blocks;

[0083] If the generation time difference of adjacent preset micro-blocks is less than the set low time difference interval, the generation difficulty of the signal is increased.

[0084] If the generation time difference of adjacent preset micro-blocks is greater than the set high time difference interval, the generation difficulty of the signal is reduced.

[0085] Compare and analyze the frequency of occurrences of signals indicating increased difficulty in the current new block with the frequency of occurrences of signals indicating decreased difficulty, and determine the difference in signal frequency and the initial adjustment direction;

[0086] If the difference in the number of signals is not greater than the set threshold for the difference in the number of signals, then the difficulty target for the current new block will not be adjusted.

[0087] If the difference in the number of signal occurrences is greater than a set threshold, the current new block will be marked as a potentially abnormal block.

[0088] The historical difficulty-related data and historical difficulty adjustment records of the potentially abnormal blocks within a preset time period are input into the difficulty-related data prediction model for prediction, thereby obtaining the potentially difficult-related data.

[0089] If no data related to the potential difficulty exceeds the corresponding preset safety threshold, the difficulty target of the potentially abnormal block is adjusted based on the difference in the number of signals and the initial adjustment direction.

[0090] If there are any difficulty-related data that may exceed the corresponding preset safety threshold range, then the corresponding difficulty adjustment measures will be matched and regarded as reference adjustment measures;

[0091] Based on the aforementioned reference adjustment measures, the predicted adjustment direction for the difficulty target of the potentially anomalous blocks is determined;

[0092] When the initial adjustment direction and the predicted adjustment direction are consistent, the difficulty target of the potentially abnormal block is adjusted based on the reference adjustment measures;

[0093] When the initial adjustment direction and the predicted adjustment direction are inconsistent, the reference adjustment measures are adjusted based on the difference in the number of signal occurrences, and then the difficulty target of the possible abnormal block is adjusted.

[0094] In the above embodiments, difficulty-related data refers to various data indicators related to the difficulty of block generation in the blockchain network, including but not limited to the total network computing power, orphan block rate, and transaction throughput. The preset security threshold range is a critical range set in advance based on the historical operating data of the blockchain network, industry experience, and security requirements to determine whether the difficulty-related data is within the normal range and whether it will pose a threat to the stability and security of the blockchain network. When the difficulty-related data exceeds this threshold range, it is considered that there may be extreme abnormalities in the blockchain network, such as large-scale network attacks or a sharp drop in computing power, so that the difficulty target can be quickly adjusted to ensure the normal operation of the blockchain network.

[0095] In the above embodiments, the safety comparison result refers to the comparison result between the real-time monitored difficulty-related data and the preset safety threshold range. It can clearly indicate whether the difficulty-related data exceeds the safety threshold, thereby providing a basis for whether to trigger difficulty adjustment measures in the future.

[0096] In the above embodiments, the difficulty adjustment measures refer to a series of operations and methods used to adjust the difficulty target of each new block in the blockchain when there is difficulty-related data that exceeds the preset security threshold range. For example, increasing or decreasing the difficulty target value to adjust the difficulty for the node computing power to find a hash value that meets the difficulty requirements.

[0097] In the above embodiments, the difficulty target refers to a parameter in the blockchain network used to specify the difficulty conditions that a node's computing power must meet when generating a new block. It determines how easy or difficult it is for the node's computing power to find a hash value that meets specific difficulty requirements through hash operations. The smaller the difficulty target value, the harder it is to find the required hash value, and the greater the difficulty of block generation; conversely, the larger the difficulty target value, the easier it is to generate a block.

[0098] In the above embodiments, the difficulty adjustment record refers to the data set generated during the process of adjusting the difficulty target, including but not limited to the adjustment time, the difficulty target value before adjustment, the difficulty target value after adjustment, and the reason for adjustment (such as which difficulty-related data and safety comparison results are used as the basis).

[0099] In the above embodiments, the preset partitioning ratio is a pre-defined ratio based on blockchain network performance optimization, difficulty adjustment accuracy, and actual application scenario requirements and design, which divides the current new block according to certain rules. By partitioning the new block according to this ratio, multiple micro-blocks can be obtained. For example, a new block can be divided into 10 micro-blocks on average.

[0100] In the above embodiments, a microblock refers to a small block obtained by dividing the current new block according to a preset division ratio. A microblock is a subdivided unit of the new block. By analyzing data such as the generation time difference of microblocks, a more detailed understanding of the blockchain network's operating status and performance can be obtained, providing a more accurate basis for difficulty adjustment. The preset number of microblocks is a pre-set number of adjacent microblocks used for analyzing time differences. The generation time difference refers to the absolute value interval between the generation times of adjacent preset number of microblocks. If the generation time difference is too small, it indicates that the network is producing blocks too quickly, posing potential security risks or performance problems; if the generation time difference is too large, it indicates network congestion or insufficient computing power.

[0101] In the above embodiments, for example, if the absolute value of the generation time difference between three adjacent micro-blocks belonging to the same new block 1 is less than the set low time difference interval, then the generation difficulty signal is increased.

[0102] For example, if the absolute value of the generation time difference between three adjacent microblocks belonging to the same new block 2 is greater than the set high time difference interval, then the generation difficulty signal is reduced.

[0103] In the above embodiments, the low time difference interval is a pre-set threshold value used to determine whether the time difference between the generation of adjacent preset micro-blocks is too small, for example, 0.5 seconds. When the time difference between the generation of adjacent preset micro-blocks is all less than this set low time difference interval, it indicates that the block generation speed of the blockchain network may be too fast, and a difficulty increase signal needs to be generated to increase the difficulty of block generation and ensure the stability and security of the network. The difficulty increase signal is used to characterize the blockchain network's need to increase the difficulty target of the current new block.

[0104] In the above embodiment, a high time difference interval is set as a pre-defined threshold value to determine whether the time difference between the generation of adjacent preset micro-blocks is too large, for example, 3 seconds. When the time difference between the generation of adjacent preset micro-blocks is greater than this set high time difference interval, it indicates that the block generation speed of the blockchain network may be too slow, and there may be network congestion or insufficient computing power. A difficulty reduction signal needs to be generated. The difficulty reduction signal is used to characterize the blockchain network's need to reduce the difficulty target of new blocks in order to speed up block generation and improve the network's transaction processing capacity.

[0105] In the above embodiments, the number of times the difficulty increase signal occurs refers to the number of times the difficulty increase signal is generated per unit time within the monitoring time range of the current new block (the monitoring time range can be set according to actual needs); the number of times the difficulty decrease signal occurs refers to the number of times the difficulty decrease signal is generated per unit time within the monitoring time range of the current new block.

[0106] In the above embodiment, for example, within the monitoring time range, whenever it is found that the generation time difference of the current new block 2 is less than the set low time difference interval, a difficulty increase signal will be generated; a total of 8 difficulty increase signals are generated within the monitoring time range; at this time, the number of times the difficulty increase signal of the current new block 2 appears is 8.

[0107] In the above embodiments, the signal frequency difference refers to the absolute difference between the number of times the difficulty increase signal appears and the number of times the difficulty decrease signal appears in the current new block; the initial adjustment direction refers to the direction of adjusting the difficulty target of the new block determined according to the signal frequency difference. When the number of times the difficulty increase signal appears is greater than the number of times the difficulty decrease signal appears, the initial adjustment direction is to increase the difficulty target; when the number of times the difficulty increase signal appears is not greater than the number of times the difficulty decrease signal appears, the initial adjustment direction is to decrease the difficulty target.

[0108] In the above embodiments, the set number difference threshold refers to the limit value of the number of signal differences used to determine whether the current new block is a potentially abnormal block.

[0109] In the above embodiments, the difficulty-related data prediction model is a pre-established model for predicting future difficulty-related data. The specific construction steps are as follows: first, collect historical difficulty-related data over a period of time; preprocess and extract features from all collected data; and train the neural network using the preprocessed data and extracted statistical features (such as the rate of change and moving average of difficulty-related data) to obtain the difficulty-related data prediction model.

[0110] In the above embodiments, the preset time period is a parameter used to determine the time range for selecting historical data; historical difficulty-related data refers to data recorded in the past period that is related to the difficulty of generating potentially abnormal blocks, including but not limited to historical difficulty target values, historical difficulty associated data, historical block generation time, etc.; historical difficulty adjustment records refer to relevant records of adjustments to the difficulty target of potentially abnormal blocks in the past period, including information such as historical adjustment time, historical difficulty target values ​​before and after historical adjustment, and historical adjustment reasons.

[0111] In the above embodiments, the possible difficulty-related data refers to the data related to the difficulty of generating future and possible abnormal blocks, which is predicted by using a difficulty-related data prediction model based on historical difficulty-related data and historical difficulty adjustment records over a preset time period.

[0112] In the above embodiments, the reference adjustment measure refers to the corresponding difficulty adjustment measure matched when there is potential difficulty-related data exceeding the corresponding preset safety threshold range; the predicted adjustment direction refers to the predicted direction for adjusting the difficulty target of potentially abnormal blocks based on the reference adjustment measure. For example, if the reference adjustment measure 1 is: the difficulty target value is reduced by 20%, then the predicted adjustment direction is a reduction in the difficulty target.

[0113] In the above embodiment, for example, within the monitoring time range of the new block 3, the number of times the difficulty increase signal appeared was 10, and the number of times the difficulty decrease signal appeared was 3. The difference in the number of signals was 7, which is greater than the set difference threshold of 5. Since 7 is greater than 5, the current new block 3 is marked as a potentially abnormal block 1; the initial adjustment direction is to increase the difficulty target.

[0114] Furthermore, since the possible difficulty-related data of the current potentially abnormal block 1 does not exceed the corresponding preset security threshold range, the difficulty target Y0 of the current potentially abnormal block is increased based on the signal count difference of 7, resulting in the adjusted difficulty target Y1 = Y0 × (7-5)ε; where ε represents the preset adjustment ratio, which is determined in advance based on the characteristics of the blockchain network, the historical difficulty adjustment effect, and security requirements.

[0115] In the above embodiments, for example, if the possible abnormal block 2 has difficulty-related data that exceeds the corresponding preset safety threshold range, then the corresponding difficulty adjustment measures are matched: the difficulty target value is increased by 8%, that is, the predicted adjustment direction is to increase the difficulty target; and within the monitoring time range, the signal frequency difference of the possible abnormal block 2 is 6, and the initial adjustment direction is to decrease the difficulty target.

[0116] The predicted adjustment direction of the possible abnormal block 2 is inconsistent with the initial adjustment direction. Based on the preset adjustment ratio ε = 2% and the signal number difference of 6 times, 8% of the reference adjustment measures are adjusted. The adjustment range of the reference adjustment measures is: (6-5)×2% = 2%.

[0117] At this point, the 8% increase in the difficulty target value in the reference adjustment measures will be reduced by 2%, that is, the adjusted increase in the difficulty target value will be: 8% - 2% = 6%, and the difficulty target value of potentially abnormal block 2 will be increased by 6%.

[0118] The beneficial effects of the above technical solution are as follows: By monitoring difficulty-related data in real time and comparing it with a preset security threshold range, extreme anomalies such as large-scale network attacks and rapid changes in computing power can be quickly detected, triggering difficulty adjustment measures in a timely manner to ensure the normal operation of the network; new blocks are divided into micro-blocks and their generation time differences are analyzed to generate signals for increasing or decreasing difficulty, providing a precise basis for difficulty adjustment; the initial adjustment direction is determined by statistically analyzing the number of signal occurrences, and possible abnormal blocks are judged by combining the set frequency difference threshold; the possible difficulty-related data of possible abnormal blocks are predicted using a difficulty-related data prediction model, and when the possible difficulty-related data exceeds the threshold, reference adjustment measures are matched to determine the predicted adjustment direction, which is analyzed in conjunction with the initial adjustment direction, so that the adjustment of the difficulty target considers both the current network state and future trends, enhancing the scientificity and rationality of the adjustment, thereby helping to improve the blockchain network's ability to cope with complex situations and ensuring the stability and security of the network.

[0119] The working principle of the above technical solution is as follows: First, the difficulty-related data is monitored in real time and compared with a preset safety threshold range. If the range is exceeded, difficulty adjustment measures are triggered to adjust the difficulty target of the new block and generate a record. If the range is not exceeded, the new block is divided into micro-blocks according to a preset ratio. Then, by comparing the time difference between adjacent preset micro-blocks with the set low time difference interval and the set high time difference interval, difficulty increase signals and difficulty decrease signals are generated. Then, the occurrence frequency of these difficulty increase and difficulty decrease signals is compared to determine the frequency difference and the initial adjustment direction. If the frequency difference is not greater than the set frequency difference threshold, the difficulty is not adjusted. The target difficulty is set; if it exceeds the threshold, it is marked as a potentially abnormal block; then, historical data and records of the potentially abnormal blocks within a preset time period are input into the prediction model to obtain potentially difficult data; if no potentially difficult data exceeds the threshold range, the difficulty target is adjusted according to the signal frequency difference and the initial adjustment direction; if potentially difficult data exceeds the threshold range, corresponding adjustment measures are matched as a reference to determine the predicted adjustment direction; finally, if the initial adjustment direction and the predicted adjustment direction are consistent, the reference measures are adjusted; if they are inconsistent, the reference measures are adjusted first according to the signal frequency difference, and then the difficulty target of the corresponding block is adjusted.

[0120] In this embodiment, a digital risk identification model for industrial transactions is constructed, and the following operations are performed:

[0121] Collect historical data on industry transactions and divide the collected historical data into training set and test set.

[0122] The deep learning model is trained using a training set, enabling it to autonomously learn the risk identification behavior of digital industrial transactions from the training set and effectively identify abnormal behaviors or potential risks in digital industrial transactions, thus establishing a deep learning-based risk identification model for digital industrial transactions.

[0123] The test set was used to perform performance testing on the deep learning-based risk identification model for digital industrial transactions, to evaluate whether the deep learning-based risk identification model for digital industrial transactions can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions, and to determine the model test evaluation results.

[0124] Based on the model testing and evaluation results, the parameters of the deep learning-based industrial transaction digital risk identification model are adjusted and optimized to determine the optimal industrial transaction digital risk identification model that can effectively identify abnormal behaviors or potential risks in industrial transaction digitalization.

[0125] In this embodiment, a test set is used to perform performance testing on the deep learning-based industrial transaction digital risk identification model, and the following operations are performed:

[0126] The test set is input into the deep learning-based digital risk identification model for industrial transactions. The risk identification performance of the deep learning-based digital risk identification model for industrial transactions is tested based on the test set, and the real-time test data of the model is determined.

[0127] The model test real-time data is compared with the preset model test threshold data to analyze the matching between the model test real-time data and the model test threshold data, and to determine whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0128] When the real-time data of the model test matches the threshold data of the model test, the deep learning-based risk warning model for digital industrial transactions can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions.

[0129] When the real-time data of the model test does not match the threshold data of the model test, the risk warning model for digital industrial transactions based on deep learning cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in the digital industrial transactions.

[0130] Specifically, the performance of the deep learning-based digital risk identification model for industrial transactions was tested using a test set. The model test evaluation results are shown in Table 1.

[0131] Table 1: Model Test Evaluation Results

[0132]

[0133] Therefore, by comparing the real-time data of model testing with the preset model testing threshold data, the matching degree between the real-time data and the model testing threshold data is analyzed, thereby determining whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in industrial transaction digitization. This facilitates the subsequent determination of the optimal industrial transaction digitization risk identification model, thereby improving the identification accuracy of abnormal behaviors or potential risks in industrial transaction digitization.

[0134] In this embodiment, determining whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behaviors or potential risks in industrial transaction digitization also includes:

[0135] When there is no real-time model test data exceeding the preset model test threshold, it is determined that the current deep learning-based industrial transaction digitization risk warning model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0136] When there is real-time model test data that exceeds the preset model test threshold, the corresponding real-time model test data will be marked as matching analysis data.

[0137] The test matching coefficient is determined by combining the data difference between the matching analysis data and the corresponding preset model test threshold data with the corresponding performance feedback weight.

[0138] When the test matching coefficient exceeds the preset matching threshold, it is determined that the current deep learning-based industrial transaction digitization risk warning model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization.

[0139] Otherwise, it can be determined that the current deep learning-based risk warning model for digital industrial transactions cannot achieve the expected effect of effectively identifying abnormal behaviors or potential risks in the digitalization of industrial transactions.

[0140] In the above embodiments, the real-time data of model testing includes categories such as accuracy, recall, false positive rate, and risk assessment accuracy. Among them, risk assessment accuracy refers to the accuracy of the model in determining the risk level of transaction samples in the test set; risk level refers to the level used by the model to characterize the degree of risk for transaction samples in the test set, including three levels: high risk, medium risk, and low risk.

[0141] In the above embodiments, the preset model test threshold data refers to the critical value set in advance for analyzing the risk warning performance of the model, based on a comprehensive consideration of factors such as the actual needs, business objectives, and expectations of the digital risk supervision and early warning system for industrial transactions; the matching analysis data refers to the real-time model test data with data values ​​greater than the corresponding preset model test threshold data.

[0142] In the above embodiments, the performance feedback weight refers to the weight value assigned according to the importance of real-time data indicators of different models in the digital risk supervision and early warning system for industrial transactions. It is used to reflect the comprehensive influence of real-time data of various models on model performance when determining the test matching coefficient. It is obtained by solving the matrix constructed by pairwise comparison and scoring using the analytic hierarchy process, and the value range is (0, 1).

[0143] In the above embodiments, the test matching coefficient is used to reflect the degree of matching between the model output and the expectation. The formula for calculating the test matching coefficient is as follows:

[0144]

[0145] In the formula, P s Represented as the test match coefficient; x iLet Δx be the i-th matching analysis data, where i = 1, 2, ..., n, and n represents the total number of matching analysis data; i This is represented as the preset model test threshold data corresponding to the i-th matching analysis data; ω1 represents the performance feedback weight for the i-th matching analysis data; ω2 represents the weight of the deviation between the matching analysis data and the preset model test threshold data that meets the model performance expectation on the recognition effect of the analysis model, with a value range of (0, 1); N represents the total number of real-time model test data; ω2 represents the weight of the proportion of matching analysis data on the recognition effect of the analysis model, with a value range of (0, 1), ω1>ω2.

[0146] The beneficial effects of the above-mentioned technical means are as follows: by comparing the real-time data of the model test with the preset model test threshold data and calculating the test matching coefficient, a quantitative basis can be provided for judging whether the deep learning-based industrial transaction digital risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitalization, which will help to determine the optimal industrial transaction digital risk identification model in the future.

[0147] In this embodiment, the parameters of the deep learning-based industrial transaction digital risk identification model are adjusted and optimized based on the model testing and evaluation results, and the following operations are performed:

[0148] When the deep learning-based risk identification model for digital industrial transactions fails to achieve the expected effect of effectively identifying abnormal behaviors or potential risks in digital industrial transactions, the parameters of the deep learning-based risk identification model for digital industrial transactions will be adjusted and optimized.

[0149] The performance of the adjusted and optimized industrial transaction digital risk identification model will be tested again to re-evaluate whether the deep learning-based industrial transaction digital risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitalization. This process will continue until the deep learning-based industrial transaction digital risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitalization, thereby determining the optimal industrial transaction digital risk identification model for effectively identifying abnormal behavior or potential risks in industrial transaction digitalization.

[0150] In this embodiment, the data of the entire industrial transaction process is analyzed and identified to determine the digital risk identification results of industrial transactions, and the following operations are performed:

[0151] The entire process of industrial transactions is input into the industrial transaction digital risk identification model. The model analyzes the entire process of industrial transactions and automatically identifies abnormal behaviors or potential risks in the digitalization of industrial transactions, thereby determining the results of the digital risk identification of industrial transactions.

[0152] The risk warning and management module is used to provide timely risk warnings based on the results of digital risk identification in industrial transactions, and to form a digital closed-loop management of industrial transactions.

[0153] In this embodiment, timely risk warnings are issued based on the results of digital risk identification in industrial transactions, and the following operations are performed:

[0154] Based on the risk identification results of digital industrial transactions, timely risk warnings are issued and the risk warning information is pushed to the management in real time. The management can then manage abnormal behaviors or potential risks in digital industrial transactions in a timely manner, including freezing some transaction permissions through smart contracts and notifying regulatory agencies to intervene.

[0155] It also tracks the entire management process and makes real-time dynamic adjustments based on the tracking feedback, forming a digital closed-loop management of industrial transactions.

[0156] In summary, by storing the entire process of industrial transactions using blockchain technology and combining it with smart contracts to automate the execution of regulatory rules, the immutability of the entire industrial transaction data is ensured. By constructing a digital risk identification model for industrial transactions, the entire process of industrial transactions is analyzed, and abnormal behaviors or potential risks in the digitalization of industrial transactions are automatically and effectively identified, thereby determining the results of digital risk identification. Based on the results of digital risk identification, timely risk warnings are issued, forming a digital closed-loop management of industrial transactions. Real-time monitoring, risk identification, and early warning handling of transaction data can be achieved based on blockchain, which can improve the effectiveness of digital management of industrial transactions.

[0157] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0158] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A blockchain-based digital risk supervision and early warning system for industrial transaction, characterized in that, Comprise: A blockchain storage module for storing industry transaction data based on blockchain technology, and realizing automatic supervision rule execution combined with smart contracts to ensure that industry transaction data cannot be tampered with; A risk supervision identification module for constructing an industry transaction digital risk identification model to analyze and identify industry transaction data and determine industry transaction digital risk identification results; A risk early warning management module for timely risk early warning based on industry transaction digital risk identification results and forming a digital closed-loop management of industry transactions; Also comprising: Real-time monitoring of difficulty correlation data and comparison with the corresponding preset safety threshold range to obtain a safety comparison result; According to the safety comparison result, if there is difficulty correlation data that exceeds the preset safety threshold range, trigger the corresponding difficulty adjustment measure to adjust the difficulty target of the new block, and generate a difficulty adjustment record; If there is no difficulty correlation data that exceeds the preset safety threshold range, divide the current new block according to the preset division ratio to obtain a microblock; When the time difference between adjacent preset microblocks is less than the set low time difference interval, a difficulty increase signal is generated; If the time difference between adjacent preset microblocks is greater than the set high time difference interval, a difficulty decrease signal is generated; Compare and analyze the number of occurrences of the difficulty increase signal and the number of occurrences of the difficulty decrease signal to determine a signal number difference value and an initial adjustment direction; If the signal number difference value is not greater than the set number difference threshold, the difficulty target of the current new block is not adjusted; If the signal number difference value is greater than the set number difference threshold, the current new block is marked as a possible abnormal block; Input the historical difficulty related data and historical difficulty adjustment record of the possible abnormal block in the preset time period into a difficulty correlation data prediction model for prediction to obtain possible difficulty related data; If there is no possible difficulty related data that exceeds the corresponding preset safety threshold range, adjust the difficulty target of the possible abnormal block based on the signal number difference value and the initial adjustment direction; If there is possible difficulty related data that exceeds the corresponding preset safety threshold range, match the corresponding difficulty adjustment measure as a reference adjustment measure; Based on the reference adjustment measure, determine the predicted adjustment direction of the difficulty target of the possible abnormal block; When the initial adjustment direction and the predicted adjustment direction are consistent, adjust the difficulty target of the possible abnormal block based on the reference adjustment measure; When the initial adjustment direction and the predicted adjustment direction are inconsistent, adjust the difficulty target of the possible abnormal block after adjusting the reference adjustment measure based on the signal number difference value; The difficulty correlation data refers to various data indicators related to the block generation difficulty in the blockchain network, including the total network computing power, the orphan block rate, and the transaction throughput; the preset security threshold range is a critical range that is set in advance based on the historical operation data of the blockchain network, industry experience, and security requirements for judging whether the difficulty correlation data is within a normal range and whether it will pose a threat to the stability and security of the blockchain network; when the difficulty correlation data exceeds this threshold range, it is considered that the blockchain network may have an extreme abnormal situation, so that the difficulty target can be quickly adjusted to ensure the normal operation of the blockchain network; The security comparison result refers to the numerical comparison result of the real-time monitored difficulty correlation data and the preset security threshold range, which can clearly indicate whether the difficulty correlation data exceeds the security threshold, thereby providing a basis for whether to trigger the difficulty adjustment measures subsequently; The difficulty adjustment measures refer to a series of operations and methods for adjusting the difficulty target of each new block in the blockchain when there is difficulty correlation data that exceeds the preset security threshold range. 2.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 1, characterized in that, Based on the blockchain technology, the whole process data of the industry transaction is stored, and the following operations are performed: Based on the smart contract, the industry transaction data is generated into data records, and the data records are packaged into blocks, with a timestamp and the hash value of the previous block; Based on the consensus algorithm, the nodes in the blockchain network verify the transaction of the new block, and after the verification, the new block is broadcast to the whole network of the blockchain, so that all nodes in the blockchain network update the ledger synchronously; The new block is added to the blockchain according to the timestamp and forms an unalterable data chain, thereby generating the whole process data of the industry transaction, wherein the data records of each transaction are queried through the blockchain browser for subsequent auditing and supervision. 3.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 2, characterized in that, Based on the consensus algorithm, the nodes in the blockchain network verify the transaction of the new block, and the following operations are performed: The block header data of the new block includes the hash value of the previous block, the timestamp of the current block, the Merkle root of the transaction data, a random number, and the difficulty target preset by the network; Based on the hash algorithm, the hash operation is performed on the block header data, wherein the random number is continuously adjusted and the hash value is recalculated until a hash value that meets the difficulty requirement is found; After the hash value that meets the difficulty requirement is found, the new block is broadcast to the whole network, and after receiving the new block, the other nodes verify the hash value, the integrity of the transaction data, and the timestamp information to ensure the validity of the new block and complete the transaction verification of the new block.

4. The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 3, characterized in that, An industry transaction digital risk identification model is constructed, and the following operations are performed: Collect industry transaction historical data and divide the collected industry transaction historical data, wherein the industry transaction historical data is divided into a training set and a test set; The training set is used to train the deep learning model, so that the deep learning model learns the industry transaction digital risk identification behavior from the training set and can effectively identify abnormal behavior or potential risks in the industry transaction digitalization, and determine the deep learning-based industry transaction digital risk identification model; The performance of the deep learning-based industrial transaction digitization risk identification model is tested by using the test set to evaluate whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization, and determine the model test evaluation result; According to the model test evaluation result, the parameters of the deep learning-based industrial transaction digitization risk identification model are adjusted and optimized to determine the optimal industrial transaction digitization risk identification model for effectively identifying abnormal behavior or potential risks in industrial transaction digitization. 5.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 4, characterized in that, The performance of the deep learning-based industrial transaction digitization risk identification model is tested by using the test set to perform the following operations: The test set is input into the deep learning-based industrial transaction digitization risk identification model, and the risk identification performance of the deep learning-based industrial transaction digitization risk identification model is tested based on the test set to determine the model test real-time data; The model test real-time data is compared with the preset model test threshold data, the matching between the model test real-time data and the model test threshold data is analyzed, and it is judged whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization; When the model test real-time data matches the model test threshold data, the deep learning-based industrial transaction digitization risk warning model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization; When the model test real-time data does not match the model test threshold data, the deep learning-based industrial transaction digitization risk warning model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization. 6.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 5, characterized in that, Judging whether the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization also includes: When there is no model test real-time data that exceeds the preset model test threshold data, it is determined that the current deep learning-based industrial transaction digitization risk warning model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization; When there is model test real-time data that exceeds the preset model test threshold data, the corresponding model test real-time data is marked as matching analysis data; By combining the data difference between the matching analysis data and the corresponding preset model test threshold data with the corresponding performance feedback weight, a test matching coefficient is determined; When the test matching coefficient exceeds the preset matching threshold, it is determined that the current deep learning-based industrial transaction digitization risk warning model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization; Otherwise, it is determined that the current deep learning-based industrial transaction digitization risk warning model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization. 7.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 4, characterized in that, According to the model test evaluation results, the parameters of the deep learning-based industrial transaction digitization risk identification model are adjusted and optimized, and the following operations are performed: When the deep learning-based industrial transaction digitization risk identification model cannot achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization, the parameters of the deep learning-based industrial transaction digitization risk identification model are adjusted and optimized. The adjusted and optimized industrial transaction digitization risk identification model is tested again, and the deep learning-based industrial transaction digitization risk identification model is re-evaluated to determine whether it can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization until the deep learning-based industrial transaction digitization risk identification model can achieve the expected effect of effectively identifying abnormal behavior or potential risks in industrial transaction digitization, and the optimal industrial transaction digitization risk identification model for effectively identifying abnormal behavior or potential risks in industrial transaction digitization is determined. 8.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 6, characterized in that, The industrial transaction full-process data is analyzed and identified to determine the industrial transaction digitization risk identification result, and the following operations are performed: The industrial transaction full-process data is input into the industrial transaction digitization risk identification model, the industrial transaction full-process data is analyzed according to the industrial transaction digitization risk identification model, and the abnormal behavior or potential risks in industrial transaction digitization are automatically and effectively identified to determine the industrial transaction digitization risk identification result. 9.The blockchain-based industrial transaction digitalization risk supervision and early warning system according to claim 8, characterized in that, According to the industrial transaction digitization risk identification result, timely risk warning is performed, and the following operations are performed: According to the industrial transaction digitization risk identification result, timely risk warning is performed, and the risk warning information is pushed to the management side in real time, the abnormal behavior or potential risks in industrial transaction digitization are timely managed by the management side, the management situation is tracked throughout the process, the real-time dynamic adjustment is performed according to the tracking feedback, and the digital closed-loop management of industrial transaction is formed.

Images