A transaction sharing method based on a delivery service platform

Abstract

The present invention relates to the field of financial technology, and specifically to a transaction sharing method based on a settlement service platform; the present invention analyzes the accuracy, timeliness and cost-effectiveness of each trader through transaction behavior data analysis and evaluation, combined with historical transaction data, so that the platform can effectively identify traders with good performance and potential risk traders; through intelligent recommendation and optimization strategies, the platform can continuously improve transaction sharing strategies to ensure that decisions are more in line with market demand and improve transaction effects; by adopting encryption algorithms and asymmetric encryption methods, transaction data including sensitive information can be effectively protected from privacy; the present invention utilizes time series analysis based on an ARIMA model, so that the platform can predict market changes in real time, take corresponding actions according to different risk levels, and send early warning information in a timely manner to help traders respond quickly.

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and specifically to a transaction sharing method based on a delivery service platform. Background Art

[0002] FinTech refers to the field of technology that uses emerging technologies to improve and innovate financial services. In many areas of financial services, trading and delivery services play a vital role. Delivery service platforms are usually used for the final settlement of transactions such as commodities and securities, and are the key link in achieving the delivery of funds and commodities (or assets) between traders. With the further development of FinTech, delivery service platforms have gradually introduced technologies such as big data analysis, blockchain, artificial intelligence, and smart contracts to optimize transaction processes, improve efficiency, reduce risks, and ensure transparency, and provide traders with more accurate transaction recommendations and support through intelligent decision-making and risk management systems.

[0003] Although financial technology has brought about tremendous changes, it still faces many challenges in its actual application.

[0004] First, data privacy protection and information security have always been the core challenges facing financial technology. With the continuous occurrence of data leaks, how to protect user privacy and ensure the security of transaction data has become a bottleneck in the development of financial technology.

[0005] Secondly, trading sharing platforms often adopt a unified sharing strategy and content, and fail to customize according to the behavioral characteristics, trading preferences and risk tolerance of different traders. The content shared by the platform is often based on the overall trend derived from big data analysis, but this shared content does not necessarily meet the specific needs of each user. For example, some traders may be more concerned with real-time alerts of market price fluctuations, while others focus on potential arbitrage opportunities or risk management. Different traders have different risk preferences and investment styles, and the existing trading sharing mechanism often fails to take these individual differences into account, resulting in the shared content not being applicable to some traders.

[0006] In view of the above problems, it is necessary to propose a transaction sharing method based on the delivery service platform. Summary of the invention

[0007] The purpose of the present invention is to solve the problems existing in the background technology and to propose a transaction sharing method based on a delivery service platform.

[0008] The purpose of the present invention can be achieved through the following technical solutions:

[0009] A transaction sharing method based on a delivery service platform comprises the following steps:

[0010] Step 1: Data collection and preparation;

[0011] Obtain the transaction data log of all participants in each transaction in the collection and delivery service platform, and obtain the data fields containing transaction data in each transaction, including buyer name BuyerName, seller name SellerName, transaction type TransactionType, quantity Quantity, price Price and timestamp Timestamp.

[0012] A hash algorithm is used to perform hash calculation on the data field to ensure that the data field is not tampered with during transmission.

[0013] Clean invalid data including missing fields and format errors, and standardize data from different sources to ensure data consistency.

[0014] Step 2: Transaction behavior data analysis and evaluation;

[0015] The trading behavior of each trader i is evaluated, and the accuracy, timeliness and cost-effectiveness of the delivery completion of each trader in all transactions are evaluated in combination with historical trading data. Where i is the trader ID, i=1,2,...,n.

[0016] For each trader i in the delivery service platform, each transaction in which he participates is numbered, and the number symbol is j; j=1, 2, ..., m.

[0017] Get the total delivery quantity in each transaction j , Correct delivery quantity 、Delivery quantity specified in the order , Actual delivery time 、Delivery time specified in the order , the total cost incurred during the transaction and the total revenue from the transaction .

[0018] By formula Calculate the trading accuracy characteristic value of each trader , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio . Generate the behavioral feature vector of each trader ( , , , ).

[0019] As a preferred embodiment of the present invention, all traders are behaviorally clustered based on behavioral feature vectors.

[0020] By formula Calculate the behavioral similarity α(i1, i2) of any two traders i1 and i2. Where c1, c2, c3 and c4 are all preset weight factors. Where i1∈i and i2∈i and i1≠i2.

[0021] The i1 and i2 in the behavior similarity calculation are traversed over all trader IDs i=1, 2, ..., n. The behavior similarity between each trader i1 and all other traders i2 is obtained. All similar traders are clustered by K-means to establish a 4-dimensional trading behavior feature space, in which the 4 dimensions correspond to the transaction accuracy feature values. , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio . Take the behavior feature vector of each trader i ( , , , ) are used as coordinates to arrange several points in the trading behavior feature space. Traders i1 and i2 that meet the following rule: the behavior similarity α (i1, i2) is less than a preset threshold are recorded as similar traders, and traders i1 and i2 are classified into the same trader cluster.

[0022] The similar traders are judged to traverse all the trader IDs i to obtain a number of trader clusters.

[0023] Calculate the center point of each trader cluster ( , , , ). Where k is the number of the trader cluster and k=1, 2, ..., K. , , and are the transaction accuracy feature values ​​of all behavioral feature vectors in the trader cluster , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio The average value over the dimension.

[0024] A plurality of virtual servers are established, the number of the virtual servers being equal to the number of trader clusters and uniquely corresponding to each trader cluster. The virtual servers are used to receive the transaction data of all traders in the trader cluster uniquely corresponding to the virtual servers.

[0025] As a preferred method of the present invention, the transaction pattern and market trend are analyzed through historical transaction data to help identify the effectiveness of the transaction strategy and analyze the potential market risks. The specific process is as follows:

[0026] Establish an ARIMA-based time series analysis model to predict future market trends and identify potential risks and market opportunities.

[0027] The time series analysis model is: .in is the predicted value of the market data at time t, including the delivery price, delivery quantity, market interest rate, transaction price volatility and delivery cost of each delivery commodity. Where a is the bias coefficient, b1, b2, ..., bw are regression coefficients; w is the preset regression time window; c is the prediction noise, and the specific value is a random value with a fluctuation range of (-ε×maxX, ε×maxX). Where ε is the preset noise weight factor, and maxX is the historical maximum value of market data X.

[0028] Based on the historical market data, the time series analysis model is fitted to solve the bias coefficient a and regression coefficients b1, b2, ..., bw.

[0029] Calculate the forecast value of market data at preset time intervals With actual value difference When the preset first risk criterion is met: When , the first-level risk signal corresponding to the market data X is output;

[0030] When the preset second risk criterion is met: When , the secondary risk signal corresponding to the market data X is output;

[0031] When the preset third risk criterion is met: When , the third-level risk signal corresponding to the market data X is output;

[0032] in is the historical data mean of market data X in time window w, is the historical data standard deviation of market data X within time window w.

[0033] Step 3: Transaction sharing and risk warning;

[0034] For each of the trader clusters, whenever a trader i in a trader cluster performs a transaction, the transaction data is sent to the central processor corresponding to the center point of the trader cluster.

[0035] As a preferred method of the present invention, a preliminary transaction sharing strategy is formulated: selectively share transaction data, encrypt sensitive data in the transaction data including the buyer's name and the seller's name through an asymmetric encryption algorithm, and share the encrypted transaction data with other traders in the same trader cluster.

[0036] As a preferred embodiment of the present invention, different risk response strategies are executed according to the first, second or third level risk signals.

[0037] Whenever a first-level risk signal corresponding to market data X is identified, it is determined that there is a slight fluctuation in the market. Although there is risk, the impact on the current transaction is small.

[0038] Whenever a secondary risk signal corresponding to market data X is identified, it is determined that the market is experiencing large fluctuations, which may affect short-term trading decisions. At this time, a market fluctuation reminder about market data X is sent to users of the delivery service platform via SMS.

[0039] Whenever a third-level risk signal corresponding to market data X is identified, it is determined that there is a significant risk in the market, and a pop-up notification is pushed in the delivery service platform's APP to ensure that users are aware of risk information in a timely manner.

[0040] Step 4: Optimize smart trading recommendations and sharing strategies;

[0041] Based on historical data and market trends, a reinforcement learning strategy model is established to iteratively optimize the initial transaction sharing strategy. Through real-time interaction between the simulated environment and users, user feedback is learned from the environment in real time to optimize the transaction sharing strategy.

[0042] The strategy update formula for reinforcement learning is: .in is the expected value of the behavior at the current moment, which indicates the expected return of executing a certain sharing strategy a under a certain market environment s. The higher the expected value, the better the effect of the current trading sharing strategy. is the expected value of the behavior at the last moment; s is a multidimensional vector, each element of which corresponds to a market data X. a is the sharing strategy matrix of n×n dimensions .

[0043] In the sharing strategy matrix, each row and column represents a trader i=1,2,...,n. The elements = , , ..., Respectively represent whether the transaction information of trader i is shared with trader i2. If yes, then The value of is 1; otherwise, The value of is 0.

[0044] The elements described The specific value of is a random number 0 or 1, and the elements on the diagonal are , , ..., The value of is always 0.

[0045] The sharing strategy matrix corresponds to a specific trading sharing scheme, which limits the trading sharing of each trader to all other traders. Where λ is the learning rate, which controls the magnitude of each expected value update. is an n-dimensional reward vector, where each element corresponds to a trader i, and its specific value represents the score of trader i for this transaction sharing; the reward vector represents the reward value obtained after executing the transaction sharing strategy; is the discount factor, representing the weight of future rewards; is the historical maximum behavioral expectation value at past moments;

[0046] Setting learning goals: maximum behavioral expectations, .

[0047] When there are u consecutive moments, the historical maximum behavior expectation value of the past moments When the specific value of converges, it is considered The corresponding transaction sharing plan is the optimal transaction sharing plan.

[0048] Step 5: Data encryption and privacy protection;

[0049] Define different roles for the delivery service platform, including administrators, traders, regulators and data analysts.

[0050] As a preferred embodiment of the present invention, corresponding access rights are assigned to each role, so that users can only access resources permitted by their roles.

[0051] Set the administrator's access rights to view all users' transaction data.

[0052] Set the trader's access rights to only view his own trading data and trading data recommended by trading sharing;

[0053] The regulator’s access rights are set to: viewing all users’ transaction data, the underlying code and data logs of the settlement service platform.

[0054] Set the data analyst access rights to view all users' transaction data and data logs.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention analyzes the accuracy, timeliness and cost-effectiveness of each trader through transaction behavior data analysis and evaluation, combined with historical transaction data, so that the platform can effectively identify traders with good performance and potential risk traders. This analysis not only helps participants refer to more dimensional behavior data when making decisions, but also provides the platform with effective risk prediction and market trend analysis. Through intelligent recommendation and optimization strategies, the platform can continuously improve transaction sharing strategies to ensure that decisions are more in line with market demand and improve transaction results.

[0057] 2. The present invention uses encryption algorithms and asymmetric encryption methods to effectively protect the privacy of transaction data, including sensitive information. In addition, the platform uses fine-grained permission management and MFA multi-factor authentication to ensure that only authorized users can access sensitive data and perform key operations. The platform sets access rights for different roles to ensure data compliance and security and reduce the risk of information leakage;

[0058] 3. The present invention uses time series analysis based on the ARIMA model. The platform can predict market changes in real time and take corresponding actions according to different risk levels. When the market fluctuates, the platform can send early warning information in time, through APP notifications or text messages, etc., to help traders respond quickly. This early warning mechanism can effectively reduce the negative impact of market fluctuations on trading decisions and improve the reaction speed and response ability of market participants. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings:

[0060] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0061] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] See also Figure 1 As shown, a transaction sharing method based on a delivery service platform comprises the following steps:

[0063] Step 1: Data collection and preparation;

[0064] Obtain the transaction data log of all participants in each transaction in the collection and delivery service platform, and obtain the data fields containing transaction data in each transaction, including buyer name BuyerName, seller name SellerName, transaction type TransactionType, quantity Quantity, price Price and timestamp Timestamp.

[0065] A hash algorithm is used to perform hash calculation on the data field to ensure that the data field is not tampered with during transmission.

[0066] Clean invalid data including missing fields and format errors, and standardize data from different sources to ensure data consistency.

[0067] Step 2: Transaction behavior data analysis and evaluation;

[0068] The trading behavior of each trader i is evaluated, and the accuracy, timeliness and cost-effectiveness of the delivery completion of each trader in all transactions are evaluated in combination with historical trading data. Where i is the trader ID, i=1,2,...,n.

[0069] For each trader i in the delivery service platform, each transaction in which he participates is numbered, and the number symbol is j; j=1, 2, ..., m.

[0070] Get the total delivery quantity in each transaction j , Correct delivery quantity 、Delivery quantity specified in the order , Actual delivery time 、Delivery time specified in the order , the total cost incurred during the transaction and the total revenue from the transaction .

[0071] By formula Calculate the trading accuracy characteristic value of each trader , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio . Generate the behavioral feature vector of each trader ( , , , ).

[0072] Furthermore, all traders are behaviorally clustered based on the behavioral feature vectors.

[0073] By formula Calculate the behavioral similarity α(i1, i2) of any two traders i1 and i2. Where c1, c2, c3 and c4 are all preset weight factors. Where i1∈i and i2∈i and i1≠i2.

[0074] The i1 and i2 in the behavior similarity calculation are traversed over all trader IDs i=1, 2, ..., n. The behavior similarity between each trader i1 and all other traders i2 is obtained. All similar traders are clustered by K-means to establish a 4-dimensional trading behavior feature space, in which the 4 dimensions correspond to the transaction accuracy feature values. , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio . Take the behavior feature vector of each trader i ( , , , ) are used as coordinates to arrange several points in the trading behavior feature space. Traders i1 and i2 that meet the following rule: the behavior similarity α (i1, i2) is less than a preset threshold are recorded as similar traders, and traders i1 and i2 are classified into the same trader cluster.

[0075] The similar traders are judged to traverse all the trader IDs i to obtain a number of trader clusters.

[0076] Calculate the center point of each trader cluster ( , , , ). Where k is the number of the trader cluster and k=1, 2, ..., K. , , and are the transaction accuracy feature values ​​of all behavioral feature vectors in the trader cluster , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio The average value over the dimension.

[0077] A plurality of virtual servers are established, the number of the virtual servers being equal to the number of trader clusters and uniquely corresponding to each trader cluster. The virtual servers are used to receive the transaction data of all traders in the trader cluster uniquely corresponding to the virtual servers.

[0078] Furthermore, historical trading data is used to analyze trading patterns and market trends to help identify the effectiveness of trading strategies. The specific process is as follows:

[0079] Establish an ARIMA-based time series analysis model to predict future market trends and identify potential risks and market opportunities.

[0080] The time series analysis model is: .in is the predicted value of the market data at time t, including the delivery price, delivery quantity, market interest rate, transaction price volatility and delivery cost of each delivery commodity. Where a is the bias coefficient, b1, b2, ..., bw are regression coefficients; w is the preset regression time window; c is the prediction noise, and the specific value is a random value with a fluctuation range of (-ε×maxX, ε×maxX). Where ε is the preset noise weight factor, and maxX is the historical maximum value of market data X.

[0081] Based on the historical market data, the time series analysis model is fitted to solve the bias coefficient a and regression coefficients b1, b2, ..., bw.

[0082] Calculate the forecast value of market data at preset time intervals With actual value difference When the preset first risk criterion is met: When , the first-level risk signal corresponding to the market data X is output;

[0083] When the preset second risk criterion is met: When , the secondary risk signal corresponding to the market data X is output;

[0084] When the preset third risk criterion is met: When , the third-level risk signal corresponding to the market data X is output;

[0085] in is the historical data mean of market data X in time window w, is the historical data standard deviation of market data X within time window w.

[0086] Step 3: Transaction sharing and risk warning;

[0087] For each of the trader clusters, whenever a trader i in a trader cluster performs a transaction, the transaction data is sent to the central processor corresponding to the center point of the trader cluster.

[0088] Furthermore, a preliminary transaction sharing strategy is formulated: selectively share transaction data, encrypt sensitive data in the transaction data including the buyer's name and the seller's name through an asymmetric encryption algorithm, and share the encrypted transaction data with other traders in the same trader cluster.

[0089] Furthermore, different risk response strategies are implemented according to the first, second or third level risk signals.

[0090] Whenever a first-level risk signal corresponding to market data X is identified, it is determined that there is a slight fluctuation in the market. Although there is risk, the impact on the current transaction is small.

[0091] Whenever a secondary risk signal corresponding to market data X is identified, it is determined that the market is experiencing large fluctuations, which may affect short-term trading decisions. At this time, a market fluctuation reminder about market data X is sent to users of the delivery service platform via SMS.

[0092] Whenever a third-level risk signal corresponding to market data X is identified, it is determined that there is a significant risk in the market, and a pop-up notification is pushed in the delivery service platform's APP to ensure that users are aware of risk information in a timely manner.

[0093] Step 4: Optimize smart trading recommendations and sharing strategies;

[0094] Based on historical data and market trends, a reinforcement learning strategy model is established to iteratively optimize the initial transaction sharing strategy. Through real-time interaction between the simulated environment and users, user feedback is learned from the environment in real time to optimize the transaction sharing strategy.

[0095] The strategy update formula for reinforcement learning is: .in is the expected value of the behavior at the current moment, which indicates the expected return of executing a certain sharing strategy a under a certain market environment s. The higher the expected value, the better the effect of the current trading sharing strategy. is the expected value of the behavior at the last moment; s is a multidimensional vector, each element of which corresponds to a market data X. a is the sharing strategy matrix of n×n dimensions .

[0096] In the sharing strategy matrix, each row and column represents a trader i=1,2,...,n. The elements = , , ..., Respectively represent whether the transaction information of trader i is shared with trader i2. If yes, then The value of is 1; otherwise, The value of is 0.

[0097] The elements described The specific value of is a random number 0 or 1, and the elements on the diagonal are , , ..., The value of is always 0.

[0098] The sharing strategy matrix corresponds to a specific trading sharing scheme, which limits the trading sharing of each trader to all other traders. Where λ is the learning rate, which controls the magnitude of each expected value update. is an n-dimensional reward vector, where each element corresponds to a trader i, and its specific value represents the score of trader i for this transaction sharing; the reward vector represents the reward value obtained after executing the transaction sharing strategy; is the discount factor, representing the weight of future rewards; is the historical maximum behavioral expectation value at past moments;

[0099] Setting learning goals: maximum behavioral expectations, .

[0100] When there are u consecutive moments, the historical maximum behavior expectation value of the past moments When the specific value of converges, it is considered The corresponding transaction sharing plan is the optimal transaction sharing plan.

[0101] Step 5: Data encryption and privacy protection;

[0102] Define different roles for the delivery service platform, including administrators, traders, regulators and data analysts.

[0103] Furthermore, corresponding access rights are assigned to each role, restricting users to access only the resources permitted by their role.

[0104] Set the administrator's access rights to view all users' transaction data.

[0105] Set the trader's access rights to only view his own trading data and trading data recommended by trading sharing;

[0106] The regulator’s access rights are set to: viewing all users’ transaction data, the underlying code and data logs of the settlement service platform.

[0107] Set the data analyst access rights to view all users' transaction data and data logs.

[0108] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure 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 collections thereof.

[0109] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and the claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations;

[0110] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A transaction sharing method based on a delivery service platform, characterized in that: The steps include: Step 1: Data collection and preparation; Obtain the transaction data log of all participants in the collection and delivery service platform in each transaction, and obtain the data field containing transaction data in each transaction; Step 2: Transaction behavior data analysis and evaluation; Evaluate the trading behavior of each trader and obtain historical trading data of different traders, including their total delivery quantity, correct delivery quantity, order-specified delivery quantity, actual delivery time, order-specified delivery time, total costs incurred during the transaction, and total revenue from the transaction in each transaction; According to the operation of historical transaction data, the transaction accuracy characteristic value, transaction recall rate, transaction time characteristic value and cost-effectiveness ratio of each trader are obtained, and the behavior characteristic vector of each trader is generated according to the transaction accuracy characteristic value, transaction recall rate, transaction time characteristic value and cost-effectiveness ratio of each trader; The behavior similarity of any two traders is obtained by numerical calculation of the behavior feature vector, and clustering is performed based on the behavior similarity to obtain several trader clusters; Establish a four-dimensional transaction behavior feature space containing the transaction accuracy feature value, transaction recall rate, transaction timeliness feature value and cost-effectiveness ratio, and arrange corresponding points in the space according to the behavior feature vector of each trader; Calculate the behavioral similarity between traders, and classify traders whose similarity is lower than a preset threshold into the same trader cluster, thereby obtaining multiple trader clusters; Calculate the average eigenvalues ​​of each trader cluster in four dimensions and use these average eigenvalues ​​as the center point; Establish an equal number of virtual servers according to the number of trader clusters. Each virtual server uniquely corresponds to a trader cluster and is responsible for receiving the trading data of all traders in the cluster. Analyze trading patterns and market trends through historical trading data to help identify the effectiveness of trading strategies and analyze potential market risks; Step 3: Transaction sharing and risk warning; Formulate a preliminary transaction sharing strategy: selectively share transaction data, encrypt sensitive data in the transaction data, including the buyer's name and the seller's name, through an asymmetric encryption algorithm, and share the encrypted transaction data with other traders in the same trader cluster; implement different risk response strategies based on level 1, 2 or 3 risk signals; Step 4: Optimize smart trading recommendations and sharing strategies; Based on historical data and market trends, a reinforcement learning strategy model is established to iteratively optimize the initial transaction sharing strategy; through real-time interaction between the simulated environment and users, user feedback is learned from the environment in real time to optimize the transaction sharing strategy; Step 5: Data encryption and privacy protection; Define different roles for the delivery service platform and assign corresponding access rights to each role.

2. A transaction sharing method based on a settlement service platform according to claim 1, characterized in that: The specific process of obtaining the behavioral similarity of any two traders through numerical calculation of the behavioral feature vector is as follows: Get the transaction accuracy characteristic value of trader i1 , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio ; Get the transaction accuracy characteristic value of trader i2 , Transaction Recall Rate , transaction time characteristic value and cost-effectiveness ratio ; By formula Calculate the behavioral similarity α(i1, i2) of any two traders i1 and i2; where c1, c2, c3 and c4 are all preset weight factors; where i1∈i and i2∈i and i1≠i2; The i1 and i2 in the behavior similarity calculation are traversed over all trader IDs i=1, 2, ..., n; Get the behavioral similarity between each trader i1 and all other traders i2.

3. A transaction sharing method based on a delivery service platform according to claim 1, characterized in that: The specific process of analyzing trading patterns and market trends through historical trading data is as follows: Establish an ARIMA-based time series analysis model to predict future market trends and identify potential risks and market opportunities; The time series analysis model is: ;in is the predicted value of the market data at time t, including the delivery price, delivery quantity, market interest rate, transaction price volatility and delivery cost of each delivery commodity; a is the bias coefficient, b1, b2, ..., bw are regression coefficients; w is the preset regression time window; c is the prediction noise, and the specific value is a random value with a fluctuation range of (-ε×maxX, ε×maxX); ε is the preset noise weight factor, and maxX is the historical maximum value of market data X; Fit the time series analysis model based on historical market data and solve the bias coefficient a and regression coefficients b1, b2, ..., bw; Calculate the forecast value of market data at preset time intervals With actual value difference ; When the preset first risk criterion is met: When , the first-level risk signal corresponding to the market data X is output; When the preset second risk criterion is met: When , the secondary risk signal corresponding to the market data X is output; When the preset third risk criterion is met: When , the third-level risk signal corresponding to the market data X is output; in is the historical data mean of market data X in time window w, is the historical data standard deviation of market data X within time window w.

4. A transaction sharing method based on a delivery service platform according to claim 1, characterized in that: The specific process of implementing different risk response strategies based on level 1, 2 or 3 risk signals is as follows: Whenever a level 1 risk signal corresponding to market data X is identified, it is determined that there is a slight fluctuation in the market. Although there is risk, the impact on the current transaction is small. Whenever a secondary risk signal corresponding to market data X is identified, it is determined that the market is experiencing large fluctuations, which may affect short-term trading decisions; at this time, a market fluctuation reminder about market data X is sent to users of the settlement service platform via SMS; Whenever a third-level risk signal corresponding to market data X is identified, it is determined that there is a significant risk in the market, and a notification is pushed through a pop-up window in the delivery service platform's APP to ensure that users are aware of risk information in a timely manner.

5. A transaction sharing method based on a delivery service platform according to claim 1, characterized in that The reinforcement learning strategy model is specifically: Set the policy update formula for reinforcement learning: ;in is the expected value of the behavior at the current moment, which indicates the expected return of executing a certain sharing strategy a under a certain market environment s. The higher the expected value, the better the effect of the current trading sharing strategy. is the expected value of the behavior at the previous moment; Where s is a multidimensional vector, each element of which corresponds to a market data X; where a is a sharing strategy matrix of n×n dimensions ; In the sharing strategy matrix, each row and column represents a trader i=1,2,...,n; the elements = , , ..., Respectively represent whether to share the transaction information of trader i with trader i2; if so, then The value of is 1; otherwise, The value of is 0; The elements described The specific value of is a random number 0 or 1, and the elements on the diagonal , , ..., The value of is always 0; The sharing strategy matrix corresponds to a specific trading sharing scheme, which limits the situation in which each trader shares trading with all other traders; where λ is the learning rate, which controls the magnitude of each expected value update; where is an n-dimensional reward vector, where each element corresponds to a trader i, and its specific value represents the score of trader i for this transaction sharing; the reward vector represents the reward value obtained after executing the transaction sharing strategy; is the discount factor, representing the weight of future rewards; It is the historical maximum behavioral expectation value at past moments.

6. A transaction sharing method based on a settlement service platform according to claim 5, characterized in that: The specific process of optimizing the transaction sharing strategy is as follows: Setting learning goals: maximum behavioral expectations, ; When there are u consecutive moments, the historical maximum behavior expectation value of the past moments When the specific value of converges, it is considered The corresponding transaction sharing plan is the optimal transaction sharing plan.

7. A transaction sharing method based on a settlement service platform according to claim 1, characterized in that: The different roles of the delivery service platform include: Administrators, traders, regulators and data analysts.

8. A transaction sharing method based on a settlement service platform according to claim 1, characterized in that: The permissions for different roles include: Set the administrator's access rights to: view all users' transaction data; Set the trader's access rights to only view his own trading data and the trading data recommended by trading sharing; The supervisor's access rights are set to: view all users' transaction data, the underlying code and data logs of the settlement service platform; Set the data analyst access rights to view all users' transaction data and data logs.

Images