Arbitration evidence construction method and device, arbitration evidence processing method and device and storage medium
By extracting multidimensional transaction data features and ownership credibility assessment indicators through federated learning methods, an arbitration evidence package is constructed, which solves the problem of low arbitration reliability in existing technologies and achieves more reliable arbitration evidence construction and arbitration determination.
Patent Information
- Application Number
- CN202511388049.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-26
- Publication Date
- 2025-11-18
AI Technical Summary
Existing methods for constructing arbitration evidence do not fully consider a variety of factors, resulting in low reliability of arbitration.
A federated learning approach is used to extract multidimensional transaction data features. Combined with multiple ownership credibility assessment indicators and transaction data hash values, an arbitration evidence package is constructed, including a feature extraction module, a hash value generation module, and an indicator quantification module.
It improves the reliability of the construction of arbitration evidence and the reliability of arbitration, and accurately reflects the authenticity and credibility of the transaction by comprehensively considering multiple factors.
Smart Images
Figure CN120973795A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of computer, and particularly relates to a method for constructing arbitration evidence, an arbitration evidence processing method and device and a storage medium. BACKGROUND
[0002] With the development of society, multi-source connection transactions are increasingly common, covering complex scenarios such as e-commerce platform transactions, Internet of Things device data interaction and social media associated transactions. This transaction mode greatly expands the business boundaries, but also causes many ownership disputes, which puts high requirements on the construction of arbitration evidence.
[0003] The existing arbitration evidence construction method usually uses a single index such as simple transaction amount and event to evaluate the credibility of the transaction, thereby constructing arbitration evidence. However, this method does not comprehensively consider various factors affecting the construction of arbitration evidence, and cannot construct reliable arbitration evidence, resulting in low arbitration reliability. SUMMARY
[0004] The present application provides a method for constructing arbitration evidence, an arbitration evidence processing method and device, and a storage medium, which can solve the technical problem that the prior art does not comprehensively consider various factors affecting the construction of arbitration evidence, cannot construct reliable arbitration evidence, and results in low arbitration reliability.
[0005] The present application provides a method for constructing arbitration evidence, comprising:
[0006] Obtaining original transaction data of multiple data sources, and extracting federal learning features of the original transaction data by using a federal learning method;
[0007] Generating corresponding transaction data hash values based on the original transaction data;
[0008] Determining a plurality of ownership credibility evaluation indexes, and determining a dynamic quantitative score of a current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes include order timestamp and historical transaction dispute rate;
[0009] According to the federal learning features, the transaction data hash values and the dynamic quantitative score, an arbitration evidence package is constructed.
[0010] Further, the federal learning features of the original transaction data are extracted by using a federal learning method, comprising:
[0011] The original transaction data is subjected to format unification processing and deduplication processing to obtain preprocessed data;
[0012] federated learning features in the preprocessed data by using a federated learning method, wherein the federated learning features include transaction correlation features, identity consistency features, behavior pattern features and data timing features.
[0013] Further, the dynamic quantitative score of the current transaction is determined according to the ownership trustworthiness evaluation index, including:
[0014] The expression of the dynamic quantitative score of the current transaction is as follows:
[0015]
[0016] Wherein S is the dynamic quantitative score of the current transaction, n is the total number of the ownership trustworthiness evaluation index, w i is the weight of the i th ownership trustworthiness evaluation index, f i (x i ) is the quantitative function corresponding to the i th ownership trustworthiness evaluation index x i .
[0017] Further, the quantitative function includes an order timestamp quantitative function and a historical transaction dispute rate transaction function, wherein the expression of the order timestamp quantitative function is as follows:
[0018]
[0019] Wherein f 订单时间戳 (x) is the order timestamp quantitative function, x is the actual order timestamp, and t0 is the expected value of the order timestamp.
[0020] The expression of the historical transaction dispute rate quantitative function is as follows:
[0021] f 历史交易纠纷率 (x) = 1-x
[0022] Wherein f 历史交易纠纷率 (x) is the historical transaction dispute rate quantitative function, and x is the historical transaction dispute rate of the participant.
[0023] Further, the determination of the weight of the ownership trustworthiness evaluation index includes:
[0024] The weight of the ownership trustworthiness evaluation index is dynamically adjusted according to the original transaction data by using an attention mechanism.
[0025] The application also provides an arbitration evidence processing method, including:
[0026] Receiving an arbitration request, wherein the arbitration request includes an arbitration evidence package constructed by the arbitration evidence construction method described above;
[0027] An original hash value is calculated according to the original transaction data, and when the original transaction data hash value in the arbitration evidence package is consistent with the original hash value, arbitration judgment is performed according to the dynamic quantitative score and the federated learning feature.
[0028] Further, the arbitration request is received, including:
[0029] A plurality of candidate arbitration nodes are selected, and the reputation of each candidate arbitration node is calculated according to the historical arbitration accuracy and response speed of the candidate arbitration node.
[0030] The candidate arbitration node with the highest reputation is selected as the final arbitration node, and the arbitration request is received through the final arbitration node.
[0031] Further, the arbitration judgment is performed according to the dynamic quantitative score and the federated learning feature, including:
[0032] When the dynamic quantitative score is equal to or higher than a preset threshold, it is determined that the arbitration is passed.
[0033] When the dynamic quantitative score is lower than the preset threshold, it is determined whether the federated learning feature is abnormal, if not, it is determined that the arbitration is passed, and if there is an abnormality, it is determined that the arbitration is not passed.
[0034] The present application also provides an arbitration evidence construction device, comprising:
[0035] The feature extraction module is used for acquiring original transaction data of multiple data sources, and extracting federated learning features of the original transaction data by using a federated learning method.
[0036] The hash value generation module is used for generating corresponding transaction data hash values based on the original transaction data.
[0037] The index quantification module is used for determining a plurality of ownership credibility evaluation indexes, and determining a dynamic quantitative score of the current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes include order timestamps and historical transaction dispute rates.
[0038] The evidence package construction module is used for constructing an arbitration evidence package according to the federated learning features, the transaction data hash values and the dynamic quantitative score.
[0039] The present application also provides a terminal device, comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, when the computer program is executed by the processor, the arbitration evidence construction method as described above is realized.
[0040] The application further provides a computer readable storage medium, comprising: a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located performs the arbitration evidence construction method as described above.
[0041] By implementing the application, the following beneficial effects are achieved:
[0042] The application extracts multi-dimensional transaction data features of transaction data by adopting a federated learning method, further calculates a dynamic quantitative score of a transaction by determining a plurality of ownership credibility evaluation indexes, and constructs arbitration evidence by combining the multi-dimensional transaction data features, the dynamic quantitative score and a transaction data hash value, thereby comprehensively considering a plurality of factors affecting arbitration, and effectively improving the reliability of arbitration evidence construction and arbitration reliability.
[0043] Further, the application can comprehensively and accurately reflect the real credibility of a transaction by comprehensively considering a plurality of ownership credibility evaluation indexes and dynamically quantifying a score, thereby effectively improving the reliability of arbitration evidence construction. BRIEF DESCRIPTION OF DRAWINGS
[0044] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.
[0045] Figure 1 is a flowchart of a construction method of arbitration evidence provided by an embodiment of the application;
[0046] Figure 2 is a flowchart of an arbitration evidence processing method provided by an embodiment of the application;
[0047] Figure 3 is a structural diagram of a construction device of arbitration evidence provided by an embodiment of the application. DETAILED DESCRIPTION
[0048] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of this application; the use of the terms "including," "comprising," or "having" and variations thereof herein is intended to be broad and encompass the terms "consisting of" and "consisting essentially of" and variations thereof. Unless otherwise required by context, singular terms shall include pluralities and vice versa. Unless otherwise required by context, the use herein of the singular is also to be construed as a use of the plural and vice versa.
[0050] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise explicitly specified.
[0051] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily independent or alternative embodiments to each other. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0052] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.
[0053] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two), and similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).
[0054] In the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the technical terms "mounting", "connecting", "connecting", "fixing" and the like should be broadly understood, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanical connection, or it can be electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the embodiments of the present application can be understood according to the specific circumstances.
[0055] Reference is made to Figure 1To solve the technical problem that the prior art does not comprehensively consider various factors affecting the construction of arbitration evidence, cannot construct reliable arbitration evidence, and leads to low arbitration reliability, an embodiment of the present application provides a construction method of arbitration evidence, comprising:
[0056] S1, obtaining original transaction data of multiple data sources, and extracting a federated learning feature of the original transaction data by using a federated learning method;
[0057] In the embodiment of the present application, the multiple data sources can include an e-commerce platform data source, an Internet of Things device data source, and a social media data source. The original transaction data of the multiple data sources obtained in the embodiment of the present application is multi-type heterogeneous data. In the embodiment of the present application, the federated learning method is a distributed machine learning method that allows multiple clients (such as mobile devices, local servers, or distributed databases) to collaboratively train a shared machine learning model while maintaining the privacy and locality of data. By using the federated learning method, cross-platform feature extraction can be performed without sharing the original transaction data, so as to obtain multi-dimensional transaction data features (i.e., federated learning features) of the original transaction data, thereby effectively protecting data privacy and reducing data security risks.
[0058] In the embodiment of the present application, the zero-knowledge proof technology can also be combined to verify the authenticity of the data without exposing the specific content.
[0059] S2, generating a corresponding transaction data hash value based on the original transaction data;
[0060] In the embodiment of the present application, the hash value corresponding to the original transaction data can be calculated as part of the arbitration evidence.
[0061] S3, determining a plurality of ownership credibility evaluation indexes, and determining a dynamic quantitative score of the current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes include an order timestamp and a historical transaction dispute rate;
[0062] In the embodiment of the present application, the plurality of ownership credibility evaluation indexes can be determined from multiple dimensions, including a transaction dimension, a participant dimension, and an environment dimension. The indexes of the transaction dimension can include an order timestamp, payment consistency, and logistics track matching degree, etc. The indexes of the participant dimension can include a historical transaction dispute rate. The environment dimension can include device fingerprints, geographical location stability, or network behavior characteristics, etc.
[0063] For the order timestamp, if the order time is too different from the expected reasonable timestamp such as the normal sales cycle of the product or the promotion activity time, there may be an anomaly;
[0064] For payment consistency, if the payment amount does not match the product price, or the payment method is abnormal (such as frequent changes in payment account, etc.), there may be abnormalities;
[0065] The matching degree of the actual logistics track of the commodity and the preset route, the expected arrival time, etc. If the logistics information is stagnant for a long time or the route is disordered, there may be abnormalities.
[0066] For the historical transaction dispute rate, if the historical transaction dispute rate of a merchant is as high as 30%, compared with a merchant with a dispute rate of only 5%, the credibility of the current transaction will be reduced.
[0067] The present application determines a plurality of ownership credibility evaluation indexes, and determines the dynamic quantitative score of the transaction according to these indexes, to evaluate the credibility of the current transaction. For example, when the order timestamp of a certain transaction is not abnormal, the logistics track matching degree is high, and the historical transaction dispute rate is low, the greater the dynamic quantitative score value calculated, the higher the credibility of the transaction, which can provide reliable data support for subsequent arbitration determination.
[0068] For the device fingerprint, it can be used to identify the uniqueness of the user's transaction device. If multiple different device fingerprints appear in the same transaction, there may be security risks;
[0069] For the geographical location stability, if the transaction occurs frequently, such as jumping from place A to place B in a short time, and the distance between the two places is far, there may be abnormalities;
[0070] For network behavior characteristics, if there are abnormal network request frequencies, use of proxy IP, etc., there may be abnormalities.
[0071] S4, according to the federated learning feature, the transaction data hash value and the dynamic quantitative score, an arbitration evidence package is constructed.
[0072] In the embodiment of the present application, the federated learning feature, the transaction data hash value and the dynamic quantitative score are integrated into a data package to form an arbitration evidence package.
[0073] The embodiment of the present application extracts multi-dimensional transaction data features of transaction data by using the federated learning method, and further calculates the dynamic quantitative score of the transaction by determining a plurality of ownership credibility evaluation indexes. The arbitration evidence is constructed by combining the multi-dimensional transaction data features, the dynamic quantitative score and the transaction data hash value, and a plurality of factors affecting arbitration are considered comprehensively, so as to effectively improve the reliability of the arbitration evidence construction and the arbitration reliability.
[0074] In one embodiment, step S1, the federated learning feature of the original transaction data is extracted by using the federated learning method, including:
[0075] S11, perform format unification processing and deduplication processing on the original transaction data to obtain preprocessed data;
[0076] S12, extract federal learning features in the preprocessed data by using a federal learning method, wherein the federal learning features include transaction correlation features, identity consistency features, behavior pattern features, and data time sequence features.
[0077] In the embodiment of the present application, the transaction correlation features are used to determine the correlation between the user's purchase behavior and the logistics delivery timeliness, the identity consistency features are used to determine the user's multi-platform identity, the behavior pattern is used to determine whether the user's cross-platform behavior pattern is consistent, and the data time sequence features are used to determine the matching degree of multi-source data time sequence. The behavior pattern is the behavior characteristics of the user on different platforms, for example, a user usually shops at night, and the behavior pattern of the user is a night shopping pattern. If a large transaction is made at an abnormal time such as early morning, it means that the transaction is inconsistent with the user's regular behavior pattern. Correlation refers to the connection or dependency relationship between different transactions, for example, if a user purchases a product on an e-commerce platform, the purchase behavior should have a direct correlation with the subsequent logistics delivery behavior (such as delivery, transportation, and delivery). If the correlation between these behaviors is abnormal, such as the time or content of the purchase behavior and the logistics delivery behavior do not match, there may be an anomaly. The multi-source data time sequence is the order and consistency of data from different data sources in time, for example, the order time of a transaction should be earlier than the payment time and the logistics delivery time. If the time sequence of the data is chaotic or inconsistent, it may indicate that the data has been tampered with or there are other problems.
[0078] The embodiment of the present application can determine whether there is a transaction anomaly through federal learning features, for example, through the identity consistency feature to determine that the user's multi-platform identity is inconsistent, which is an identity consistency anomaly; through the transaction correlation feature to determine that the correlation between the user's purchase behavior and the logistics delivery timeliness is abnormal, which is a transaction correlation anomaly, and so on.
[0079] The embodiment of the present application extracts federal learning features in the preprocessed data by using a federal learning method, which can extract multi-dimensional transaction data features without sharing the original transaction data, effectively protecting data privacy and reducing data security risks. Moreover, the multi-dimensional transaction data features are used to construct an arbitration evidence package, which comprehensively considers arbitration-related data, effectively improving arbitration reliability.
[0080] In one embodiment, step S3, determining a dynamic quantitative score of the current transaction according to the ownership credibility evaluation index, includes:
[0081] The expression of the dynamic quantitative score of the current transaction is as follows:
[0082]
[0083] wherein S is a dynamic quantification score of the current transaction, n is the total number of the weight-attribute credibility evaluation indexes, w i is the weight of the i-th weight-attribute credibility evaluation index, f i (x i ) is the quantification function corresponding to the i-th weight-attribute credibility evaluation index x i .
[0084] In the embodiments of the present application, by comprehensively integrating the weight-attribute credibility evaluation indexes in multiple dimensions and dynamically quantifying the score, the real credibility of the transaction can be comprehensively and accurately reflected, thereby effectively improving the reliability of the arbitration evidence construction.
[0085] In one embodiment, the quantification function includes an order timestamp quantification function and a historical transaction dispute rate transaction function, wherein the expression of the order timestamp quantification function is as follows:
[0086]
[0087] wherein f 订单时间戳 (x) is the order timestamp quantification function, x is the actual order timestamp, and t0 is the expected value of the order timestamp.
[0088] In the embodiments of the present application, the expected order timestamp range can be set according to historical data, and the middle value of the range is taken as the expected value of the order timestamp. The order timestamp quantification function can map the order timestamp to the interval of 0-1, and the closer to the expected reasonable timestamp, the closer to 1 the score is, and vice versa. Assuming that the expected reasonable range of the order timestamp of a certain commodity is 10:00-12:00, t0 is 11:00, and if the order timestamp x is 11:10, the quantification score will be close to 1 when substituted into the formula, indicating that the order timestamp credibility is high.
[0089] The expression of the historical transaction dispute rate quantification function is as follows:
[0090] f 历史交易纠纷率 (x)=1-x
[0091] wherein f 历史交易纠纷率 (x) is the historical transaction dispute rate quantification function, and x is the historical transaction dispute rate of the participant.
[0092] In the embodiments of the present application, the lower the historical transaction dispute rate, the higher the score of the historical transaction dispute rate quantification function. For example, if the historical transaction dispute rate x of a certain merchant is 0.1, the quantification score is 0.9, which is high, indicating that the merchant has good reputation.
[0093] In one embodiment, the determination of the weight of the weight-attribute credibility evaluation index includes:
[0094] An attention mechanism is adopted to dynamically adjust the weight of the ownership trustworthiness evaluation index according to the original transaction data.
[0095] In the embodiment of the present application, it can be determined according to the original transaction data whether there is a logistics track abnormal situation, if yes, the weight of the logistics data in the evaluation can be dynamically increased through the attention mechanism. For example, the logistics information shows that the package stays at a certain location for too long, and the weight of the logistics track matching degree index is automatically increased through the attention mechanism, so that it occupies a larger proportion in the comprehensive evaluation.
[0096] The embodiment of the present application can identify and highlight the key features in the transaction data through the attention mechanism, so as to assign a higher weight to the important ownership trustworthiness evaluation index, effectively improve the reliability of the dynamic quantitative score determination, and further effectively improve the reliability of the arbitration evidence construction.
[0097] The embodiment of the present application has the following beneficial effects:
[0098] The embodiment of the present application extracts the multi-dimensional transaction data features of the transaction data by using the federated learning method, and further calculates the dynamic quantitative score of the transaction by determining a plurality of ownership trustworthiness evaluation indexes, and constructs the arbitration evidence by combining the multi-dimensional transaction data features, the dynamic quantitative score and the transaction data hash value, comprehensively considers a plurality of factors affecting arbitration, and thus can effectively improve the reliability of the arbitration evidence construction and the arbitration reliability.
[0099] Further, by comprehensively considering a plurality of dimensions of the ownership trustworthiness evaluation index and dynamically quantifying the score, the real trustworthiness of the transaction can be comprehensively and accurately reflected, thereby effectively improving the reliability of the arbitration evidence construction.
[0100] Please refer to Figure 2 The embodiment of the present application also provides an arbitration evidence processing method, comprising:
[0101] S10, receiving an arbitration request, the arbitration request comprising an arbitration evidence package constructed by the arbitration evidence construction method provided in the above embodiment;
[0102] S20, calculating an original hash value according to the original transaction data, and when the original transaction data hash value in the arbitration evidence package is consistent with the original hash value, arbitrating according to the dynamic quantitative score and the federated learning feature.
[0103] In the embodiment of the present application, when the original transaction data hash value in the arbitration evidence package is consistent with the original hash value, it indicates that the transaction data has not been tampered with, and on this basis, arbitration is determined according to the dynamic quantitative score and the federated learning feature, which can effectively improve the reliability of the arbitration determination.
[0104] In one embodiment, step S10, receiving an arbitration request, comprises:
[0105] S101, selecting a plurality of candidate arbitration nodes, and calculating the credibility of each candidate arbitration node according to the historical arbitration accuracy and response speed of the candidate arbitration node.
[0106] In an embodiment of the present application, the plurality of candidate arbitration nodes can include industry associations, notary organizations, technical experts and other decentralized arbitration nodes.
[0107] In an embodiment of the present application, the plurality of candidate arbitration nodes can be distributed in different geographical locations, and communicate and cooperate through PBFT consensus algorithm.
[0108] For example, when a transaction dispute is submitted to the arbitration platform, each candidate arbitration node will receive relevant information at the same time, quickly reach a consensus through the PBFT algorithm, determine the priority and processing flow of the dispute processing, and improve the arbitration efficiency.
[0109] The credibility of the candidate arbitration node is expressed as follows:
[0110]
[0111] Wherein, R is the credibility of the candidate arbitration node, A is the historical arbitration accuracy, V is the response speed, and a is the weight coefficient, the value range is [0, 1], which can be adjusted according to the actual application scene, and is used to balance the importance of arbitration accuracy and response speed in the credibility evaluation. For example, assuming that the value of a is 0.6, the historical arbitration accuracy A of a certain arbitration node is 0.9, the response speed V is 2h on average to process a dispute, Then the credibility R of the candidate arbitration node is 0.6*0.9+(1-0.6)*0.5=0.74, which is higher than the credibility of other candidate nodes, so the arbitration node with the highest credibility is preferred to participate in arbitration.
[0112] S102, the candidate arbitration node with the highest credibility is selected as the final arbitration node, and the arbitration request is received through the final arbitration node.
[0113] In an embodiment of the present application, the candidate arbitration node with the highest credibility is selected as the final arbitration node, and the arbitration request is received to process the arbitration data.
[0114] In an embodiment of the present application, the candidate arbitration node with the highest credibility is selected as the final arbitration node, and the arbitration request is received to process the arbitration data.
[0115] In one embodiment, step S20, arbitration judgment is made according to the dynamic quantitative score and the federal learning feature, comprising:
[0116] S201. When the dynamic quantification score is equal to or higher than a preset threshold, it is determined that the arbitration is passed.
[0117] In the embodiments of the present application, the preset threshold can be set according to actual needs, for example, it can be 0.6, 0.8, etc.
[0118] S202. When the dynamic quantification score is lower than the preset threshold, it is determined whether the federated learning features are abnormal, if not, it is determined that the arbitration is passed, if yes, it is determined that the arbitration is not passed.
[0119] In the embodiments of the present application, the federated learning features include transaction correlation features, identity consistency features, behavior pattern features and data time sequence features, and it can be determined whether the transaction correlation is abnormal, whether the identity consistency is abnormal, whether the behavior pattern is abnormal and whether the data time sequence is abnormal, if any of them is abnormal, it is determined that the arbitration is not passed.
[0120] The embodiments of the present application provide a quantitative evaluation standard through the dynamic quantification score, which can quickly and accurately determine the overall credibility of the transaction, when the dynamic quantification score is equal to or higher than the preset threshold, it is determined that the arbitration is passed, which can effectively improve the arbitration efficiency while ensuring the reliability of the arbitration, and when the dynamic quantification score is lower than the preset threshold, it is further combined with the federated learning features to determine whether the arbitration is passed, which effectively improves the arbitration reliability.
[0121] In one embodiment, natural language processing technology can also be combined to assist arbitration, including:
[0122] When the dynamic quantification score is lower than the preset threshold, the abnormal situation of the federated learning features is determined, the natural language processing technology is used to obtain the transaction related text for semantic analysis, and the arbitration is made by combining the semantic analysis result and the abnormal situation. For example, when dealing with the dispute between a consumer and a merchant about the quality of goods, the texts such as the evaluation content of the consumer and the response of the merchant are analyzed, and the key information such as the keywords of "poor quality" and "not consistent with the description" is extracted, which provides more accurate basis for arbitration.
[0123] In the embodiments of the present application, the arbitration result and the related evidence data can be permanently recorded through the blockchain, the distributed ledger characteristics of the blockchain ensure the data unforgeability and traceability, at the same time, a hierarchical storage architecture is adopted, the core evidence (such as the arbitration result, the key evidence hash value, etc.) is stored on the blockchain, and the original transaction data (such as a large number of transaction records, logistics information, etc.) is stored in a distributed file system (such as IPFS), which not only ensures the data security, but also reduces the storage cost.
[0124] In one embodiment, the arbitration rule can be predefined by the smart contract, the manual arbitration can be automatically started when the dynamic quantitative score is lower than a preset threshold, and the preset smart contract can be directly used for arbitration when the transaction amount is less than a certain value, such as 1000 yuan, and the evidence is complete.
[0125] The embodiment of the present application has the following beneficial effects:
[0126] When the hash value of the original transaction data in the arbitration evidence package is consistent with the original hash value, the arbitration judgment is made according to the dynamic quantitative score and the federated learning feature in the embodiment of the present application, which can ensure that the arbitration judgment is made when the transaction data is not tampered with, and can effectively improve the reliability of the arbitration judgment.
[0127] Further, the embodiment of the present application provides a quantitative evaluation standard by the dynamic quantitative score, which can quickly and accurately judge the overall credibility of the transaction, and determine that the arbitration is passed when the dynamic quantitative score is equal to or higher than a preset threshold, which can effectively improve the arbitration efficiency while ensuring the arbitration reliability, and further combine the federated learning feature to determine whether the arbitration is passed when the dynamic quantitative score is lower than the preset threshold, which effectively improves the arbitration reliability.
[0128] As shown in the above embodiment of the method for constructing arbitration evidence, corresponding device embodiments are provided; Figure 3
[0129] The embodiment of the present application provides a device for constructing arbitration evidence, which comprises:
[0130] The feature extraction module 10 is configured to acquire original transaction data of multiple data sources and extract federated learning features of the original transaction data by using a federated learning method;
[0131] The hash value generation module 20 is configured to generate corresponding transaction data hash values based on the original transaction data;
[0132] The index quantification module 30 is configured to determine a plurality of ownership credibility evaluation indexes and determine a dynamic quantitative score of the current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes comprise an order timestamp and a historical transaction dispute rate;
[0133] The evidence package construction module 40 is configured to construct an arbitration evidence package according to the federated learning features, the transaction data hash values and the dynamic quantitative score.
[0134] In one embodiment, the federated learning features of the original transaction data are extracted by using a federated learning method, which comprises:
[0135] The original transaction data is subjected to format unification processing and de-duplication processing to obtain preprocessed data;
[0136] The federated learning method is used to extract the federated learning features in the preprocessed data, wherein the federated learning features include transaction correlation features, identity consistency features, behavior pattern features and data timing features.
[0137] In one embodiment, the dynamic quantitative score of the current transaction is determined according to the ownership trustworthiness evaluation index, including:
[0138] The expression of the dynamic quantitative score of the current transaction is as follows:
[0139]
[0140] Wherein S is the dynamic quantitative score of the current transaction, n is the total number of the ownership trustworthiness evaluation index, w i is the weight of the i-th ownership trustworthiness evaluation index, f i (x i ) is the quantitative function corresponding to the i-th ownership trustworthiness evaluation index x i .
[0141] In one embodiment, the quantitative function includes an order timestamp quantitative function and a historical transaction dispute rate transaction function, wherein the expression of the order timestamp quantitative function is as follows:
[0142]
[0143] Wherein f 订单时间戳 (x) is the order timestamp quantitative function, x is the actual order timestamp, and t0 is the expected value of the order timestamp.
[0144] The expression of the historical transaction dispute rate quantitative function is as follows:
[0145] f 历史交易纠纷率 (x)=1-x
[0146] Wherein f 历史交易纠纷率 (x) is the historical transaction dispute rate quantitative function, and x is the historical transaction dispute rate of the participant.
[0147] In one embodiment, the determination of the weight of the ownership trustworthiness evaluation index includes:
[0148] The attention mechanism is used to dynamically adjust the weight of the ownership trustworthiness evaluation index according to the original transaction data.
[0149] It can be understood that the above-mentioned device item embodiment is corresponding to the method item embodiment of the present application, which can realize the construction method of the arbitration evidence provided by any one of the above-mentioned method item embodiments of the present application.
[0150] It should be noted that the apparatus embodiments described above are only illustrative, and part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. In addition, in the apparatus embodiment provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which can be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.
[0151] On the basis of the above-mentioned embodiment of the arbitration evidence construction method, another embodiment of the present application provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the arbitration evidence construction method of any one embodiment of the present application is realized.
[0152] For example, in this embodiment, the computer program can be divided into one or more modules, one or more modules are stored in the memory and executed by the processor to complete the present application. One or more module elements can be a series of computer program instruction segments capable of completing a specific function, which is used to describe the execution process of the computer program in the terminal device.
[0153] The terminal device can be a desktop computer, a notebook computer, a palm computer and a cloud server, etc. The terminal device can include, but is not limited to, a processor and a memory.
[0154] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and connects all parts of the terminal device through various interfaces and lines.
[0155] On the basis of the above-mentioned embodiment of the method, another embodiment of the present application provides a computer readable storage medium, which comprises a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located executes the arbitration evidence construction method of any one of the above-mentioned embodiments of the present application.
[0156] The modules / units integrated in the device / terminal equipment can be stored in a computer readable storage medium if they are implemented in the form of software function units and sold or used as independent products. Based on this understanding, all or part of the processes in the above-mentioned embodiments can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms. The computer readable medium can include any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.
[0157] The above is the preferred embodiment of the present application. It should be noted that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements are also considered within the scope of protection of the present application.
Claims
1. A method of constructing an arbitral evidence, characterized by, The method comprises the following steps: obtaining original transaction data of multiple data sources, and extracting federated learning features of the original transaction data by using a federated learning method; generating corresponding transaction data hash values based on the original transaction data; determining multiple ownership credibility evaluation indexes, and determining a dynamic quantitative score of a current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes comprise an order timestamp and a historical transaction dispute rate; constructing an arbitration evidence package according to the federated learning features, the transaction data hash values and the dynamic quantitative score.
2. The method of claim 1, wherein the arbitration evidence is constructed by, The step of extracting the federated learning features of the original transaction data by using the federated learning method comprises the following steps: performing format unification processing and de-duplication processing on the original transaction data to obtain preprocessed data; extracting federated learning features in the preprocessed data by using a federated learning method, wherein the federated learning features comprise transaction correlation features, identity consistency features, behavior pattern features and data time sequence features.
3. The method of claim 1, wherein the arbitration evidence is constructed by, The step of determining the dynamic quantitative score of the current transaction according to the ownership credibility evaluation indexes comprises the following steps: The expression of the dynamic quantitative score of the current transaction is as follows: where S is the dynamic quantification score of the current transaction, n is the total number of weight credibility evaluation indexes, w i is the weight of the i-th weight credibility evaluation index, f i (x i ) is the quantification function corresponding to the i-th weight credibility evaluation index x i .
4. The method of claim 3, wherein the arbitration evidence is constructed by, The quantitative function comprises an order timestamp quantitative function and a historical transaction dispute rate transaction function, wherein the expression of the order timestamp quantitative function is as follows: In the formula, f 订单时间戳 (x) is an order timestamp quantization function, x is an actual order timestamp, and t0 is an expected value of the order timestamp. The expression of the historical transaction dispute rate quantitative function is as follows: f 历史交易纠纷率 (x) = 1 - x In the formula, f 历史交易纠纷率 (x) is a historical transaction dispute rate quantification function, and x is the historical transaction dispute rate of the participant.
5. The method of arbitrating evidence construction of claim 3, wherein, The determination of the weight of the ownership credibility evaluation indexes comprises the following steps: dynamically adjusting the weight of the ownership credibility evaluation indexes according to the original transaction data by using an attention mechanism.
6. An arbitrated evidence handling method characterized by comprising: The method comprises the following steps: receiving an arbitration request, wherein the arbitration request comprises an arbitration evidence package constructed by the arbitration evidence construction method according to any one of claims 1 to 5; calculating an original hash value according to the original transaction data, and when the original transaction data hash value in the arbitration evidence package is consistent with the original hash value, performing arbitration determination according to the dynamic quantitative score and the federated learning features.
7. The arbitrated evidence processing method of claim 6, wherein, The step of receiving the arbitration request comprises the following steps: selecting multiple candidate arbitration nodes, calculating the credit of each candidate arbitration node according to the historical arbitration accuracy and response speed of the candidate arbitration nodes, and selecting the candidate arbitration node with the highest credit as a final arbitration node to receive the arbitration request. The step of performing arbitration determination according to the dynamic quantitative score and the federated learning features comprises the following steps:
8. The arbitrated evidence processing method of claim 6, wherein, when the dynamic quantitative score is equal to or higher than a preset threshold, determining that the arbitration is passed; when the dynamic quantitative score is lower than the preset threshold, determining whether the federated learning features are abnormal, and if not, determining that the arbitration is passed, and if so, determining that the arbitration is not passed. The method comprises the following steps:
9. An arbitration evidence construction apparatus characterized by comprising: a feature extraction module is configured to obtain original transaction data of multiple data sources, and extract federated learning features of the original transaction data by using a federated learning method; a hash value generation module is configured to generate corresponding transaction data hash values based on the original transaction data; a hash value generation module is configured to generate corresponding transaction data hash values based on the original transaction data; An index quantification module is configured to determine a plurality of ownership credibility evaluation indexes, and determine a dynamic quantification score of a current transaction according to the ownership credibility evaluation indexes, wherein the ownership credibility evaluation indexes include an order timestamp and a historical transaction dispute rate; An evidence package construction module is configured to construct an arbitration evidence package according to the federated learning features, the transaction data hash value, and the dynamic quantification score.
10. A computer-readable storage medium, characterized in that, The arbitration evidence construction method comprises the following steps: A stored computer program, wherein when the computer program is running, the device where the computer readable storage medium is located is controlled to execute the arbitration evidence construction method according to any one of claims 1-5.