A trusted data trading platform and method based on blockchain and big data technology

Abstract

The present invention discloses a trusted data trading platform and method based on blockchain and big data technology, relates to the field of trusted data technology, and solves the problem that risk control and user experience cannot be effectively balanced in the selection of response strategies based on risk prediction. The present invention firstly marks the collected user information, transaction amount and other data and then trains them based on a neural network model to obtain a risk prediction model that can output the risk level after inputting transaction behavior and user information. Then, by setting the corresponding risk response strategy for each level of the risk prediction model, an objective function is established with user satisfaction and risk response strategy level as influencing factors, and a risk response strategy is selected based on the condition of maximizing the objective function. In this way, while identifying the current transaction risk, it is further ensured that after identifying the transaction risk, the selected risk response strategy can satisfy both user satisfaction and risk response needs.

Description

Technical Field

[0001] The present invention relates to the field of trusted data technology, and in particular to a trusted data trading platform and method based on blockchain and big data technology. Background Art

[0002] Data transactions play a vital role in today's digital world. They not only promote the flow of information across industries and fields, but are also a key factor in promoting economic growth, innovation and competitiveness. With the development of technologies such as big data, artificial intelligence and cloud computing, data transactions have become an important means for enterprises to gain insights, optimize business processes and develop new business models. In this process, the importance of risk prediction has become increasingly prominent. It not only involves the security of financial transactions, but also includes personal privacy protection, compliance supervision and the preservation and appreciation of data assets. Effective risk prediction can help enterprises and organizations identify potential risk points in advance, so as to take preventive measures, reduce losses and ensure the security and reliability of transactions. In every link of data transactions, from data collection, processing, storage to analysis and sharing, risk prediction is the basis for ensuring data integrity, availability and confidentiality, and plays an irreplaceable role in maintaining the trust of enterprises and customers, protecting corporate reputation and market position.

[0003] However, current risk prediction technologies still have significant shortcomings. In particular, after risk prediction, the choice of response strategies is often too single and lacks flexibility and adaptability. These strategies usually do not fully consider user feedback and ignore the maximization of user satisfaction. At the same time, they do not take the maximization of safety performance as a core goal, which may result in an inability to effectively balance risk control and user experience in actual operations. In addition, the existing risk response strategy selection steps often lack comprehensive consideration of multiple factors. This limitation may lead to suboptimal selection of risk management measures and may even trigger new risks. Therefore, there is an urgent need to develop more advanced and comprehensive risk management technologies to achieve more accurate and dynamic risk prediction and response.

[0004] In view of this, a trusted data trading platform and method based on blockchain and big data technology is needed. Summary of the invention

[0005] In view of the problem that the existing technology cannot effectively balance risk control and user experience in the selection of response strategies based on risk prediction, the present invention provides a trusted data trading platform and method based on blockchain and big data technology, which can establish and solve the objective function after predicting the risk level, with the goals of maximizing user satisfaction, minimizing risk-strategy differences and maximizing security levels, and then select the most appropriate risk response strategy. The specific technical solution is as follows:

[0006] A trusted data trading platform based on blockchain and big data technology, including:

[0007] A data supply unit, connected to a data supplier and used for accessing data from the data supplier;

[0008] A data receiving unit, connected to a data demander, and used to transmit data to the data demander;

[0009] A security verification unit is connected to the data supply unit and the data receiving unit. The security verification unit includes a risk warning module and a security matching module. The risk warning module is provided with a risk warning model. The risk warning model is obtained by collecting data including transaction records and user information and marking the data with risk levels based on deep neural network training. The model is used to predict the risk level of the data supplier and the data demander in the data transaction process. The output result of the risk warning model is stored on the blockchain. The security matching module is preset with risk response strategies of different levels corresponding to the risk level. After the risk warning module outputs the risk level, the security matching module selects the risk response strategy according to the risk level.

[0010] Among them, the security matching module establishes an objective function with user satisfaction and risk response strategy level as influencing factors, and selects risk response strategies based on the maximization of the objective function. The objective function is expressed as follows:

[0011] ;

[0012] In the formula, U is user satisfaction, which is quantified by the historical ratings of user feedback. Users include at least data suppliers and data demanders; S is the risk response strategy level.

[0013] Preferably, the influencing factors of the objective function also include a risk-strategy difference, where the risk-strategy difference is the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. In this case, the objective function is expressed as follows:

[0014] ;

[0015] In the formula, U is user satisfaction, which is obtained through the scores of feedback from data suppliers and data demanders; S is the risk response strategy level; D is the risk-strategy difference;

[0016] The constraints of the objective function are as follows:

[0017] User satisfaction constraints:

[0018] ;

[0019] in, is the minimum acceptable value of user satisfaction;

[0020] Security level constraints:

[0021] ;

[0022] in, is the minimum acceptable value of the safety level;

[0023] Risk level constraints:

[0024] ;

[0025] in, R is the risk level, is the maximum acceptable value of the risk level;

[0026] Risk-Strategy Difference Constraints:

[0027] ;

[0028] in, is the target risk level, It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

[0029] Preferably, the values ​​of user satisfaction are as follows:

[0030] ;

[0031] In the formula, U 1 is the average of all users’ historical ratings of the strategy, U 2 is the average of the historical ratings of the strategy by the current trading users; and .

[0032] Preferably, the security matching module includes a supplier security matching module and a demander security matching module, both of which are connected to the risk warning module, and the risk response strategy set in the supplier security matching module is only for the data supplier, and the user satisfaction of the objective function in the supplier security matching module only uses the historical score of the data supplier; the risk response strategy set in the demander security matching module is only for the data demander, and the user satisfaction of the objective function in the demander security matching module only uses the historical score of the data demander.

[0033] Preferably, the training process of the risk warning model is as follows:

[0034] Collect data, including user information, transaction amount, transaction time, transaction location, payment method, login time, login location, login device, login IP address, and number of abnormal login attempts;

[0035] After preprocessing the data, including removing missing values ​​and standardizing, the risk levels are converted into numerical labels and the data are labeled accordingly;

[0036] Based on the neural network training model, an initial learning rate is set, and training is performed based on the labeled data to eventually obtain a trained risk warning model, which is used to output the risk level of the current transaction.

[0037] Preferably, it also includes a data transaction security unit, which includes a data desensitization module, a data sandbox module, a privacy computing module and a metering module;

[0038] Among them, the data desensitization module transforms personal identity information, financial records, etc. according to preset desensitization rules and policies; the data sandbox module provides a safe sandbox environment to ensure that data can only be calculated in the sandbox environment and the calculation results are returned. The original data cannot leave the sandbox environment, so that data can be destroyed on demand. It is used by the data supply and demand parties to complete data calculation without leaking data and algorithms to each other when one party provides the algorithm and the other party provides the data, and deliver the data to the party that needs the data results; the privacy computing module supports the original data of the cooperative organization to not leave the organization, and the participants exchange parameters through encryption mechanism, build a virtual computing model without violating data privacy protection regulations, and only share data calculation results; the measurement module cleans the data according to the demands of the data demander.

[0039] A trusted data transaction method based on blockchain and big data technology, comprising the following steps:

[0040] Collect data including transaction behaviors and user information, pre-process the data and label the data with risk levels, and derive a risk warning model based on deep neural network training. Transaction behaviors include login time, login location, login device, login IP address, transaction amount, and payment method, and user information includes user ID and geographic location.

[0041] Set corresponding risk response strategies for each risk level;

[0042] Obtain user information and transaction behaviors of data suppliers and data demanders in real time, and output the risk level of the current transaction based on the risk warning model;

[0043] The objective function is established with user satisfaction and risk response strategy level as influencing factors, and the risk response strategy is selected based on the maximization of the objective function. The objective function is expressed as follows:

[0044] ;

[0045] In the formula, U is user satisfaction, which is quantified by the historical ratings of user feedback. Users include at least data suppliers and data demanders; S is the risk response strategy level.

[0046] Preferably, the following steps are also included: the influencing factor of the objective function also includes the risk-strategy difference, and the risk-strategy difference is the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. In this case, the objective function is expressed as follows:

[0047] ;

[0048] In the formula, U is user satisfaction, which is obtained through the scores of feedback from data suppliers and data demanders; S is the risk response strategy level; D is the risk-strategy difference;

[0049] The constraints of the objective function are as follows:

[0050] User satisfaction constraints:

[0051] ;

[0052] in, is the minimum acceptable value of user satisfaction;

[0053] Security level constraints:

[0054] ;

[0055] in, is the minimum acceptable value of the safety level;

[0056] Risk level constraints:

[0057] ;

[0058] in, R is the risk level, is the maximum acceptable value of the risk level;

[0059] Risk-Strategy Difference Constraints:

[0060] ;

[0061] in, is the target risk level, It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

[0062] A computer-readable storage medium, the computer-readable storage medium comprising a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the trusted data transaction method based on blockchain and big data technology as described above.

[0063] A processor is used to run a program, wherein when the program is running, the trusted data transaction method based on blockchain and big data technology as described above is executed.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] The present invention firstly collects user information, transaction amount, transaction time, transaction location, payment method, login time, login location, login device, login IP address, abnormal login attempt number and other data, annotates them and then trains them based on a neural network model, and obtains a risk prediction model that can output the risk level after inputting the transaction behavior and user information. Then, by setting the corresponding risk response strategy for each level of the risk prediction model, an objective function is established with user satisfaction and risk response strategy level as influencing factors, and a risk response strategy is selected based on the maximization of the objective function. In this way, while identifying the current transaction risk, it is further ensured that after identifying the transaction risk, the selected risk response strategy can satisfy both the user's satisfaction and the risk response needs.

[0066] The influencing factors of the objective function set by the present invention include user satisfaction, risk response strategy level and risk-strategy difference. That is to say, the present invention will first consider the user's feedback for the selection of risk response strategy, quantify the user's satisfaction through the user's historical feedback score, and then ensure that the risk response strategy finally selected can reflect the user's satisfaction, that is to say, the final risk response strategy selection takes into account the user's subjective will. Secondly, the influencing factors of the objective function include the risk response strategy level, that is, the present invention will be more inclined to say a higher level of risk response strategy level in the selection of risk response strategy to ensure a higher level of transaction security. Finally, the influencing factors of the objective function also include risk-strategy difference, which reflects the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. This is mainly to take into account the problem of risk response cost. A higher level of risk response strategy means a higher cost while having a higher degree of security. Therefore, it is necessary to set the risk-strategy difference to constrain the selection of risk level, so as to avoid the problem that the level of risk response strategy is too high and does not match the actual demand. In addition, the present invention also adds a weight before each influencing factor, and the weight can be flexibly adjusted according to the actual situation and the nature of the current type of transaction to ensure the flexibility of the objective function. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0068] Figure 1 It is the system principle diagram of the present invention;

[0069] Figure 2 is a flow chart of the method of the present invention;

[0070] Figure 3 This is the data processing flow chart in the data sandbox module. DETAILED DESCRIPTION

[0071] 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 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.

[0072] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0073] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0074] It should be further understood that the term "and / or" used in the present description and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0075] In one embodiment of the present invention, a trusted data trading platform based on blockchain and big data technology is provided. Figure 1 As shown, including:

[0076] A data supply unit, used for accessing data from a data supplier;

[0077] A data receiving unit, used for transmitting data to a data demander;

[0078] A safety verification unit, including a risk warning module and a safety matching module, wherein the risk warning module is provided with a trained risk warning model;

[0079] The training process of the risk warning model is as follows:

[0080] S1: Data collection, obtaining training sets.

[0081] The training set can be obtained by:

[0082] Internal data: data extracted from the transaction database of financial institutions, including but not limited to users' transaction records, account information, historical transaction behaviors, etc.

[0083] External data: This may include public fraud case data, industry risk reports, economic indicators, etc., which can be obtained through public data sets, partners or third-party data providers.

[0084] The final training set includes:

[0085] User information: user ID, age, gender, occupation, geographic location, etc. Users include both data providers and data recipients.

[0086] Transaction details: transaction amount, transaction time, transaction location, payment method, transaction type (online / offline), transaction frequency, etc.

[0087] Account behavior: login time, login location, login device, login IP address, number of abnormal login attempts, etc.

[0088] Historical risk records: historical fraud records, complaint records, account freezing records, etc.

[0089] Transaction context: the time when the transaction takes place (e.g., holiday, weekday), the reputation of the transaction location, etc.

[0090] S2: Data preprocessing:

[0091] Data cleaning: remove missing values, outliers and duplicate records.

[0092] Feature engineering: construct new features, such as extracting time periods of the day, working days / non-working days of the week, etc. from transaction times.

[0093] Data encoding: One-hot encoding or label encoding for categorical variables.

[0094] Data standardization: standardize or normalize numerical features.

[0095] Label processing: Convert risk levels into numerical labels, for example, no risk is 0 and risk levels are 1 to 5.

[0096] S3: Get the training set, train the model based on the neural network, and finally get the trained risk warning model, which is used to output the risk level of the current transaction. In this process, you can set an initial learning rate, such as 0.01. And minimize the objective function (such as cross entropy loss or mean square error) through optimization algorithms (such as gradient descent). Finally, you can use the validation set to evaluate the performance of the model and perform hyperparameter tuning.

[0097] The following is a specific example to illustrate the model training process:

[0098] Suppose a financial institution wants to improve the capabilities of its transaction monitoring system to identify different levels of transaction risk. Here is a detailed description of each step:

[0099] Data Collection:

[0100] Extract users' transaction records and account behavior data from the database of financial institutions.

[0101] Extract users' abnormal login attempts and historical fraud records from security logs.

[0102] Data preprocessing:

[0103] Use the Pandas library to preprocess the data, including data cleaning and feature engineering. Create new features, such as extracting the time of day from the trading time.

[0104] Model training:

[0105] Use the random forest model in the Scikit-learn library to train the data; set the initial learning rate to 0.1; train the model and select the best model parameters through cross-validation.

[0106] Finally, the user’s transaction characteristics and account behavior characteristics are input into the trained model, and the risk level score of the transaction is output, ranging from 0 (no risk) to 5 (highest risk).

[0107] Through this model, financial institutions can identify the different risk levels of transactions in real time and take corresponding risk control measures, such as adding verification steps, limiting transaction amounts or notifying users, thereby improving transaction security and protecting users' funds.

[0108] Risk response strategies are pre-set and stored in the security matching module, and the risk response strategies include adding verification steps, restricting transactions, user notifications, risk investigations, etc.

[0109] Specifically, the additional verification step can be multi-factor authentication, including requiring users to perform secondary verification, such as sending a verification code to the user's mobile phone or email, or using biometric technology (fingerprint, facial recognition, etc.). It can also be manual review, including setting up an automatic process to submit the transaction to the manual review team for further review.

[0110] Restrictions on transactions can be amount limits, including limiting the amount of transactions a user can make, especially for new or high-risk users. They can also be transaction speed limits, including slowing down transactions, for example, by setting a cooling-off period to prevent large transactions in quick succession.

[0111] User notifications can be real-time notifications, such as: when a high-risk transaction is detected, the user is notified immediately via SMS, email or in-app to confirm whether it is a personal operation. It can also be a security reminder to provide security advice to users, such as changing passwords regularly and being wary of phishing websites.

[0112] Risk investigation includes abnormal behavior analysis, such as in-depth analysis of users' transaction behavior to determine whether there are abnormal patterns or signs of fraud. It also includes linked account review, such as checking other accounts related to high-risk transactions to identify potential fraud gangs.

[0113] In addition, risk response strategies are stored by level, that is, different risk levels correspond to different risk response strategies, and risk response strategies can be used independently or in combination. As shown in Table 1:

[0114] Table 1 Risk registration-risk response strategy one-to-one correspondence table

[0115] Risk Level Risk Response Strategies 1 User Notifications 2 User Notification + Risk Investigation 3 User notification + additional verification steps …… …… 5 Restricted Trading

[0116] After the risk warning module outputs the risk level, different risk response strategies are adopted according to different risk levels.

[0117] In one embodiment of the present invention, the data supply unit, the data receiving unit and the risk warning module are the same as those described above, except for the security matching module. In this embodiment, the security matching module is pre-set and stored with risk response strategies, which include adding verification steps, restricting transactions, user notifications, risk investigations, etc. The risk response strategies are stored by level, that is, different risk levels correspond to different risk response strategies, and the risk response strategies can be used independently or in combination. In addition, one risk level can correspond to a variety of risk response strategies for selection, as shown in Table 2:

[0118] Table 2 Risk registration-risk response strategy many-to-one correspondence table

[0119] Risk level (risk strategy level) Risk Response Strategies 1 User Notifications 1 Add verification steps 1 User Notification + Risk Investigation …… …… 2 User notification + risk investigation + additional verification steps …… ……

[0120] The problem of selecting a risk response strategy will arise here. In this embodiment, the most appropriate risk response strategy is selected by establishing an objective function and solving it, with the goals of maximizing user satisfaction, minimizing risk-strategy differences, and maximizing security levels.

[0121] The objective function is defined as follows:

[0122]

[0123] Where U is user satisfaction, which is obtained through the user feedback score; S is the security level (risk response strategy level). The security level here is quantified by the risk level. The significance is that the risk response strategy with a higher risk level has a better effect. D is the risk-strategy difference. This parameter reflects the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. For example, the current risk level is 1, and the risk response strategy adopted is "user notification + risk investigation + additional verification steps" (its corresponding risk level is 2), then the difference between them is 1.

[0124] The values ​​of user satisfaction are as follows:

[0125] ;

[0126] In the formula, U 1 is the average of all users’ historical ratings of the strategy, U 2 is the average of the current user's historical ratings of the strategy; and .

[0127] For this objective function, formulate the following constraints:

[0128] User satisfaction constraints:

[0129] ;

[0130] in, It is the minimum acceptable value of user satisfaction.

[0131] Security level constraints:

[0132] ;

[0133] in, is the minimum acceptable value for the security level.

[0134] Risk level constraints:

[0135] ;

[0136] in,R is the risk level, is the maximum acceptable value of the risk level.

[0137] Risk-Strategy Difference Constraints:

[0138]

[0139] in, is the target risk level, It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

[0140] The weight values ​​and the maximum or minimum acceptable values ​​mentioned above can be specifically set by technical experts and other personnel according to actual conditions, and this embodiment does not make any specific limitations.

[0141] Solve the objective function and obtain the best risk response strategy. The algorithm for solving the problem is an existing technology and will not be elaborated here. The overall idea is to input the user satisfaction, risk-strategy difference and security level maximization of each strategy, and output a risk response strategy with the highest objective function value.

[0142] The security verification unit limits the data transaction process of the data supply unit and the data receiving unit according to the risk response strategy output by the security matching module.

[0143] In one embodiment of the present invention, in addition to the data supply unit, data receiving unit and security verification unit mentioned above, a data transaction security unit is also provided, including a data desensitization module, a data sandbox module, a privacy computing module and a measurement module; wherein the data desensitization module transforms personal identity information, financial records, etc. according to preset desensitization rules and policies; the data sandbox module ensures that sensitive data does not leave the domain through operations such as encrypted transmission and private key decryption; the privacy computing module ensures that privacy information is not leaked during the calculation process; the measurement module cleans the data according to the requirements of the data demander.

[0144] Call the data desensitization module to deform the data through desensitization rules such as randomization and suppression, so as to achieve reliable protection of sensitive data and ensure the normal operation of the business system without leaking data. For example, when partial data samples are required, desensitized data can be provided, and some fields of partial data can be desensitized as supply data.

[0145] Call the data sandbox module to provide a safe sandbox environment, ensure that data can only be calculated in the sandbox environment, and return the calculation results. The original data cannot leave the sandbox environment, so that data can be destroyed on use. It is used by the data supply and demand parties to complete data calculation without leaking data and algorithms to each other when one party provides the algorithm and the other party provides the data, and deliver the data to the party that needs the data results. Figure 3 As shown, the data processing steps in the data sandbox module are:

[0146] The data demander transfers the business model to the data sandbox hosted by the data manager;

[0147] The data provider first generates a random data encryption key locally, encrypts the local data with the data encryption key, and then encrypts the data encryption key with the sandbox's public key;

[0148] The data provider transmits the encrypted data encryption key and encrypted data to the data sandbox hosted by the data manager;

[0149] After receiving the business model and encrypted data, the data sandbox first uses the private key to decrypt the encrypted data encryption key. After obtaining the data encryption key, it decrypts the encrypted data to obtain the original data.

[0150] After the data sandbox decrypts the original data, it runs the business model based on the original data to perform calculations and obtain calculation results;

[0151] The data manager will perform sensitive data detection on the calculation results. If they do not contain sensitive data, they will be returned to the data requester. Otherwise, an error will be returned to ensure that sensitive data does not leave the domain.

[0152] The privacy computing module is called to support the original data of the cooperative organization to remain within the organization. The participants exchange parameters through an encryption mechanism, build a virtual computing model without violating data privacy protection regulations, and only share data calculation results for data confidentiality scenarios.

[0153] In one embodiment of the present invention, a trusted data transaction method based on blockchain and big data technology is provided. Figure 2 As shown, the following steps are included:

[0154] Collect data including transaction behaviors and user information, pre-process the data and label the data with risk levels, and derive a risk warning model based on deep neural network training. Transaction behaviors include login time, login location, login device, login IP address, transaction amount, and payment method, and user information includes user ID and geographic location.

[0155] Set corresponding risk response strategies for each risk level;

[0156] Obtain user information and transaction behaviors of data suppliers and data demanders in real time, and output the risk level of the current transaction based on the risk warning model;

[0157] The objective function is established with user satisfaction and risk response strategy level as influencing factors, and the risk response strategy is selected based on the maximization of the objective function. The objective function is expressed as follows:

[0158] ;

[0159] In the formula, U is user satisfaction, which is quantified by the historical ratings of user feedback. Users include at least data suppliers and data demanders; S is the risk response strategy level.

[0160] Preferably, the following steps are also included: the influencing factor of the objective function also includes the risk-strategy difference, and the risk-strategy difference is the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. In this case, the objective function is expressed as follows:

[0161] ;

[0162] In the formula, U is user satisfaction, which is obtained through the scores of feedback from data suppliers and data demanders; S is the risk response strategy level; D is the risk-strategy difference;

[0163] The constraints of the objective function are as follows:

[0164] User satisfaction constraints:

[0165] ;

[0166] in, is the minimum acceptable value of user satisfaction;

[0167] Security level constraints:

[0168] ;

[0169] in, is the minimum acceptable value of the safety level;

[0170] Risk level constraints:

[0171] ;

[0172] in, R is the risk level, is the maximum acceptable value of the risk level;

[0173] Risk-Strategy Difference Constraints:

[0174] ;

[0175] in, is the target risk level, It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

[0176] In summary, the present invention firstly collects user information, transaction amount, transaction time, transaction location, payment method, login time, login location, login device, login IP address, abnormal login attempt number and other data, annotates them and then trains them based on a neural network model, and obtains a risk prediction model that can output the risk level after inputting transaction behavior and user information. Then, by setting corresponding risk response strategies for each level of risk prediction model, an objective function is established with user satisfaction and risk response strategy level as influencing factors, and a risk response strategy is selected based on the maximization of the objective function. In this way, while identifying the current transaction risk, it is further ensured that after identifying the transaction risk, the selected risk response strategy can satisfy both user satisfaction and risk response needs.

[0177] In addition, the influencing factors of the objective function set by the present invention include user satisfaction, risk response strategy level and risk-strategy difference. That is to say, the present invention will first consider the user's feedback for the selection of risk response strategy, quantify the user's satisfaction through the user's historical feedback score, and then ensure that the risk response strategy finally selected can reflect the user's satisfaction, that is to say, the final risk response strategy selection takes into account the user's subjective will. Secondly, the influencing factors of the objective function include the risk response strategy level, that is, the present invention will be more inclined to say a higher level of risk response strategy level in the selection of risk response strategy, so as to ensure a higher level of transaction security. Finally, the influencing factors of the objective function also include risk-strategy difference, which reflects the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. This is mainly to take into account the problem of risk response cost. A higher level of risk response strategy means a higher cost while having a higher degree of security. Therefore, it is necessary to set the risk-strategy difference to constrain the selection of risk level, so as to avoid the problem that the level of risk response strategy is too high and does not match the actual demand. In addition, the present invention also adds a weight before each influencing factor, and the weight can be flexibly adjusted according to the actual situation and the nature of the current type of transaction, so as to ensure the flexibility of the objective function.

[0178] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0179] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0180] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0181] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A trusted data trading platform based on blockchain and big data technology, characterized in that: include: A data supply unit, connected to a data supplier and used for accessing data from the data supplier; A data receiving unit, connected to a data demander, and used to transmit data to the data demander; A security verification unit is connected to the data supply unit and the data receiving unit. The security verification unit includes a risk warning module and a security matching module. The risk warning module is provided with a risk warning model. The risk warning model is obtained by collecting data including transaction records and user information and marking the data with risk levels based on deep neural network training. It is used to predict the risk level of the data supplier and the data demander in the data transaction process. The output result of the risk warning model is stored on the blockchain; the security matching module is preset with risk response strategies of different levels corresponding to the risk level. After the risk warning module outputs the risk level, the security matching module selects the risk response strategy according to the risk level; the risk response strategy includes a single one of adding verification steps, restricting transactions, user notifications, and risk investigations, or a combination of more than two, and the risk response strategy is stored by level classification, and one risk level corresponds to a plurality of risk response strategies for selection; Among them, the security matching module establishes an objective function with user satisfaction and risk response strategy level as influencing factors, and selects risk response strategies based on the maximization of the objective function. The objective function is expressed as follows: Maximize Z=ω1·U+ω2·S-ω3·D; In the formula, U is user satisfaction, which is quantified by the historical ratings of user feedback. Users include at least data suppliers and data demanders; S is the level of risk response strategy. The higher the level of risk response strategy used, the better the effect; D is the risk-strategy difference, which is the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision; The constraints of the objective function include user satisfaction constraints, which are as follows: U≥U min ; Among them, U min is the minimum acceptable value of user satisfaction; The constraints of the objective function also include risk-strategy difference constraints, as follows: |R-R target |≤R diff_max Among them, R target is the target risk level, R diff_max It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision; The values ​​of user satisfaction are as follows: U=k1U1+k2U2; Where U1 is the mean of all users’ historical ratings of the strategy, and U2 is the mean of the current trading users’ historical ratings of the strategy; k1 and k2 are the weight factors of U1 and U2 respectively, and k2>k1.

2. A trusted data trading platform based on blockchain and big data technology according to claim 1, characterized in that: The constraints of the objective function also include: Security level constraints: S≥S min ; Among them, S min is the minimum acceptable value of the safety level; Risk level constraints: R≤R max ; Among them, R is the risk level, R max is the maximum acceptable value of the risk level; Risk-Strategy Difference Constraints: |R-R target |≤R diff_max ; Among them, R target is the target risk level, R diff_max It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

3. A trusted data trading platform based on blockchain and big data technology according to any one of claims 1-2, characterized in that: The security matching module includes a supplier security matching module and a demander security matching module. Both the supplier security matching module and the demander security matching module are connected to the risk warning module. The risk response strategy set in the supplier security matching module is only for the data supplier, and the user satisfaction of the objective function in the supplier security matching module only uses the historical score of the data supplier; the risk response strategy set in the demander security matching module is only for the data demander, and the user satisfaction of the objective function in the demander security matching module only uses the historical score of the data demander.

4. A trusted data trading platform based on blockchain and big data technology according to claim 2, characterized in that: The training process of the risk warning model is as follows: Collect data, including user information, transaction amount, transaction time, transaction location, payment method, login time, login location, login device, login IP address, and number of abnormal login attempts; After preprocessing the data, including removing missing values ​​and standardizing, the risk levels are converted into numerical labels and the data are labeled accordingly; Based on the neural network training model, an initial learning rate is set, and training is performed based on the labeled data to eventually obtain a trained risk warning model, which is used to output the risk level of the current transaction.

5. A trusted data trading platform based on blockchain and big data technology according to claim 1, characterized in that: It also includes a data transaction assurance unit connected to the data supply unit and the data demand unit, and the data transaction assurance unit includes a data desensitization module, a data sandbox module, a privacy computing module and a metering module; Among them, the data desensitization module transforms personal identity information, financial records, etc. according to preset desensitization rules and policies; the data sandbox module provides a safe sandbox environment to ensure that data can only be calculated in the sandbox environment and the calculation results are returned. The original data cannot leave the sandbox environment, so that data can be destroyed on demand. It is used by the data supply and demand parties to complete data calculation without leaking data and algorithms to each other when one party provides the algorithm and the other party provides the data, and deliver the data to the party that needs the data results; the privacy computing module supports the original data of the cooperative organization to not leave the organization, and the participants exchange parameters through encryption mechanism, build a virtual computing model without violating data privacy protection regulations, and only share data calculation results; the measurement module cleans the data according to the demands of the data demander.

6. A trusted data transaction method based on blockchain and big data technology, applied to a trusted data transaction platform based on blockchain and big data technology as described in any one of claims 1 to 5, characterized in that: The following steps are involved: Collect data including transaction behaviors and user information, pre-process the data and label the data with risk levels, and derive a risk warning model based on deep neural network training. Transaction behaviors include login time, login location, login device, login IP address, transaction amount, and payment method, and user information includes user ID and geographic location. Set corresponding risk response strategies for each risk level; Obtain user information and transaction behaviors of data suppliers and data demanders in real time, and output the risk level of the current transaction based on the risk warning model; The objective function is established with user satisfaction and risk response strategy level as influencing factors, and the risk response strategy is selected based on the maximization of the objective function. The objective function is expressed as follows: Maximize Z = ω1·U + ω2·S; In the formula, U is user satisfaction, which is quantified by the historical ratings of user feedback. Users include at least data suppliers and data demanders; S is the risk response strategy level.

7. A trusted data transaction method based on blockchain and big data technology according to claim 6, characterized in that: The following steps are also included: The influencing factors of the objective function also include the risk-strategy difference, which is the difference between the current risk level and the level corresponding to the risk response strategy adopted in the decision. In this case, the objective function is expressed as follows: Maximize Z=ω1·U+ω2·S-ω3·D; In the formula, U is user satisfaction, which is obtained through the scores of feedback from data suppliers and data demanders; S is the risk response strategy level; D is the risk-strategy difference; The constraints of the objective function are as follows: User satisfaction constraints: U≥U min ; Among them, U min is the minimum acceptable value of user satisfaction; Security level constraints: S≥S min ; Among them, S min is the minimum acceptable value of the safety level; Risk level constraints: R≤R max ; Among them, R is the risk level, R max is the maximum acceptable value of the risk level; Risk-Strategy Difference Constraints: |R-R target |≤R diff_max ; Among them, R target is the target risk level, R diff_max It is the maximum acceptable value of the difference between the actual risk level and the risk level corresponding to the decision.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the trusted data transaction method based on blockchain and big data technology as described in any one of claims 6 to 7.

9. A processor, characterized in that: The processor is used to run a program, wherein when the program is running, the trusted data transaction method based on blockchain and big data technology described in any one of claims 6 to 7 is executed.

Images