E-commerce transaction link optimization method and system based on risk analysis

Abstract

The invention relates to the technical field of transaction links, in particular to an e-commerce transaction link optimization method and system based on risk analysis, and the method comprises the following steps: carrying out the man-machine verification of a user based on an e-commerce transaction platform, extracting the time and error times spent by the user in the man-machine verification process, and carrying out the calculation of the time and error times; and in combination with the login times and the login device type of the user, evaluating a user risk index to obtain a user risk evaluation result. By analyzing the deviation between the behavior data of the user in the verification process and the normal verification data and combining login and equipment types of the user, the risk index of the user can be effectively evaluated and quantified, the accuracy and flexibility of risk detection are improved, and the risk detection accuracy is improved by aiming at different transaction time periods and user risk levels. And the verification steps and verification contents are flexibly adjusted, so that the problems of poor user experience and low efficiency caused by excessive verification or an unreasonable verification process are reduced, and high efficiency and smoothness of transactions are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of transaction link, and in particular to an optimization method and system for an e-commerce transaction link based on risk analysis. Background Art

[0002] The technical field of transaction link mainly involves the interaction process of various links in the e-commerce platform, including product release, user order placement, payment settlement, logistics distribution, and after-sales service. This field focuses on how to improve efficiency, reduce costs, mitigate risks, enhance user experience, and ensure transaction security by optimizing each link in the transaction link. Transaction link technology usually combines various technical means, including data analysis, risk management, payment security mechanisms, supply chain management, etc.

[0003] Among them, the optimization method for e-commerce transaction link focuses on optimizing the transfer efficiency and security of each link in the e-commerce transaction process, reducing potential transaction risks and improving customer satisfaction. Its main purpose is to realize the intelligence and automation of the transaction process in the e-commerce platform, ensure the smooth progress of transactions, and reduce the negative impacts brought by delays, payment problems, or logistics bottlenecks in the link. It includes risk analysis and decision support based on data, path optimization technology, enhancement of payment security mechanisms, etc., aiming to improve the reliability and efficiency of the overall transaction chain.

[0004] The security protection measures in the traditional method lack flexibility during the transaction process, and the verification means mostly rely on fixed patterns, making it difficult to cope with the diversity of different users and transaction environments, and difficult to comprehensively consider the dynamic changes of user behavior. Moreover, when dealing with large-scale transaction data, potential high-risk behaviors cannot be identified in a timely manner, and security vulnerabilities are likely to occur. The traditional method's verification and security management of users rely on preset standard processes and lack the ability to adjust verification means according to real-time data and risk situations. As a result, during transactions with high-risk periods or high-risk users, there are no timely enhanced verification measures, which cannot effectively prevent fraud or transaction failures, increasing the security risks of the platform and affecting the user's transaction experience. Especially when facing frequent verification operations, the risk of user loss increases. Summary of the Invention

[0005] The purpose of the present invention is to solve the drawbacks existing in the prior art, and to propose an optimization method and system for an e-commerce transaction link based on risk analysis.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: An optimization method for an e-commerce transaction link based on risk analysis, comprising the following steps:

[0007] S1: Based on the e-commerce trading platform, conduct a human-machine verification on the user, extract the time spent and the number of errors made by the user during the human-machine verification process, and combine the number of login times and the type of login device of the user to evaluate the user risk index and obtain the user risk assessment result;

[0008] S2: Based on the trading records of the user on the e-commerce trading platform, compare them with the trading information of known user types to identify the user type and obtain the user type classification result;

[0009] S3: Based on the user type classification result and the user risk assessment result, according to the order information of the user on the e-commerce trading platform currently, extract the purchase data of the same type of goods by similar users, evaluate the deviation of the purchase data, and combine the user risk index to evaluate the trading risk level of the current order and obtain the order risk assessment result;

[0010] S4: Based on the trading records of the e-commerce trading platform, count the trading data of the e-commerce trading platform in each time period, and combine the standard deviation and abnormal trading frequency of the trading data in each period to evaluate the required security level for the current period and obtain the time period security analysis result;

[0011] S5: Based on the order risk assessment result and the time period security analysis result, according to the trading risk level of the order and the required security level for the current period, evaluate the required security verification level for the user, adjust the number of steps and the type of verification content in the user's security verification process, and after the security verification, implement the order transaction to obtain the order transaction verification result.

[0012] The improvement of the present invention is that the steps for obtaining the user risk assessment result are as follows:

[0013] S111: Based on the e-commerce trading platform, conduct a human-machine verification on the user, collect the data of the user during the human-machine verification process, including the verification time and the number of errors of the user, to obtain the human-machine verification record:

[0014] S112: Based on the human-machine verification record, collect the user login device information, detect whether the user login device is a commonly used device and a known device. If it is a known commonly used device, mark it as 0, otherwise mark it as 1, to obtain the user login device identification score information;

[0015] S113: Based on the user device status marking information and the human-machine verification record, through the formula:

[0016]

[0017] Calculate the user risk index to obtain the user risk assessment result;

[0018] where, J userJ is the verification time for the user during the human-machine verification process avg E is the average verification time for normal users user E is the number of errors for the user during the human-machine verification process avg L is the average number of errors for normal users R D is the number of login attempts by the user type α is the score of the user's login device identifier R β is the weight coefficient of the bias term R γ is the weight coefficient of the number of logins R is the weight coefficient of the device type

[0019] The improvement of the present invention is that the step of obtaining the user type division result is as follows:

[0020] S211: Based on the transaction records of the user on the e-commerce trading platform, extract the types of goods traded by the user, the average purchase amount, and the number of purchases to obtain the user purchase characteristic data

[0021] S212: Based on the user purchase characteristic data, compare it with the transaction information of known user types through the formula:

[0022]

[0023] Calculate the matching degree of the user type

[0024] where B is the matching degree of the user type, w i is the weight coefficient of the i-th feature, P user，i is the purchase data of the user on the i-th feature, P known，i is the purchase data of the known user type on the i-th feature, and n is the total number of features

[0025] S213: Based on the matching degree of the user type, by comparing the magnitudes of the matching degrees of each user type, select the user type with the maximum match as the matching type to obtain the user type division result

[0026] The improvement of the present invention is that the step of evaluating the deviation of the purchase data is as follows:

[0027] S311: Collect the order information of the user on the e-commerce trading platform currently, including the number of orders, the order amount, and the type of goods, to obtain the user order data

[0028] S312: Based on the user order data and the user type division result, extract the purchase data of similar users, including the average purchase quantity and the average purchase amount of similar goods, to obtain the purchase data of similar users

[0029] S313: Based on the user order data and the purchase data of similar users, through the formula:

[0030]

[0031] Calculate the purchase data deviation index, evaluate the deviation degree according to the size of the deviation index, and obtain the purchase data deviation evaluation result;

[0032] Among them, M is the purchase data deviation index, Q user is the purchase quantity of the user in the current order, Q avg is the average purchase quantity of similar users for this type of goods, A user is the purchase amount of the current user in the current order, A avg is the average purchase amount of similar users for this type of goods, μ M is the exponential coefficient.

[0033] The improvement of the present invention is that the obtaining steps of the order risk assessment result are:

[0034] S321: Based on the user risk assessment result and the purchase data deviation assessment result, according to the order risk management standard, extract the preset minimum transaction risk level and the maximum transaction risk level to obtain the transaction risk correlation data;

[0035] S322: Based on the transaction risk correlation data, through the formula:

[0036]

[0037] Calculate the transaction risk level of the current order to obtain the order risk assessment result;

[0038] Among them, L order is the transaction risk level of the current order, L min is the preset minimum transaction risk level, L max is the preset maximum transaction risk level, M is the purchase data deviation index, R is the user risk index, M max is the preset maximum purchase data deviation value, R max is the preset maximum user risk index, α L and β L are weight coefficients.

[0039] The improvement of the present invention is that the obtaining steps of the period safety analysis result are:

[0040] S411: Based on the transaction records of the e-commerce trading platform, statistically analyze the transaction data of the e-commerce trading platform in each time period, including the average number of transactions, the average total transaction amount, and the average number of transaction participants in each period, calculate the standard deviation of the transaction data in each period, and statistically analyze the abnormal transaction frequency in each period to obtain the statistical information of the transaction data in the time period;

[0041] S412: Based on the statistical information of the transaction data in the time period, through the formula:

[0042]

[0043] Calculate the data deviation coefficient of the current time period;

[0044] where S is the data deviation coefficient of the current time period, T cu is the average number of transactions in the current time period, T avg is the average number of transactions in all time periods, A cu is the average transaction amount in the current time period, A avg is the average transaction amount in all time periods, N cu is the average number of transaction participants in the current time period, N avg is the average number of transaction participants in all time periods, STD cu is the standard deviation of the transaction data in the current time period, STD avg is the standard deviation of the transaction data in all time periods, Fcu is the average abnormal transaction frequency in the current time period, F avg is the average abnormal transaction frequency in all time periods, α W β W γ W δ W ∈ W is the weight coefficient;

[0045] S413: Based on the data deviation coefficient of the current time period, through the formula:

[0046]

[0047] Calculate the required security level for the current time period to obtain the analysis result of the security level in the time period;

[0048] where L required is the required security level for the current time period, L base is the reference security level value, and S is the data deviation coefficient of the current time period.

[0049] The improvement of the present invention is that the step of obtaining the order transaction verification result is:

[0050] S511: Based on the order risk assessment result and the time period security analysis result, according to the transaction risk level of the order and the security level required for the current time period, through the formula:

[0051]

[0052] Calculate the required security verification level of the user;

[0053] where, L ver is the required security verification level of the user, L Z is the benchmark security verification level, L order is the transaction risk level of the current order, L oavg is the average transaction risk level of all orders, L required is the required security level for the current time period, L ravg is the average required security level for all time periods, α v and β v are weight coefficients;

[0054] S512: Based on the required security verification level of the user, according to the preset security verification standard, adjust the number of steps and the type of verification content in the user's security verification process, including SMS verification, fingerprint verification, and face verification. After the security verification, implement the order transaction to obtain the order transaction verification result.

[0055] An e-commerce transaction link optimization system based on risk analysis, the e-commerce transaction link optimization system based on risk analysis is used to execute the above-mentioned e-commerce transaction link optimization method based on risk analysis, and the system includes:

[0056] The user risk analysis module, based on the e-commerce transaction platform, conducts a human-machine verification on the user, extracts the time spent and the number of errors by the user during the human-machine verification process, and combines the user's login times and login device types to evaluate the user risk index and obtain the user risk assessment result;

[0057] The user type identification module, based on the transaction records of the user on the e-commerce transaction platform, compares the type of transaction goods, average purchase amount, and purchase times with the transaction information of known user types to identify the user type and obtain the user type classification result;

[0058] The order risk analysis module, based on the user type classification result and the user risk assessment result, according to the order information of the user on the e-commerce transaction platform currently, extracts the purchase data of similar users for the same type of goods, evaluates the deviation of the purchase data, and combines the user risk index to evaluate the transaction risk level of the current order and obtain the order risk assessment result;

[0059] The time period impact analysis module, based on the transaction records of the e-commerce trading platform, statistically analyzes the transaction data of the e-commerce trading platform for each time period, and combines the standard deviation and abnormal transaction frequency of the transaction data for each time period to evaluate the required security level for the current time period, obtaining the time period security analysis result;

[0060] The transaction verification implementation module, based on the order risk assessment result and the time period security analysis result, evaluates the required security verification level for the user according to the transaction risk level of the order and the required security level for the current time period, adjusts the number of steps and the type of verification content in the user's security verification process, and after the security verification, implements the order transaction to obtain the order transaction verification result.

[0061] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0062] In the present invention, by analyzing the deviation between the behavior data of the user during the verification process and the normal verification data, and combining the user's login and device type, the risk index of the user can be effectively evaluated and quantified, improving the accuracy and flexibility of risk detection. By comparing the user's transaction history, goods type, purchase amount and other information, the user type can be identified and their transaction behavior can be evaluated, providing a more accurate judgment basis for the order risk level. By flexibly adjusting the verification steps and verification content according to different transaction time periods and user risk levels, such as SMS verification, fingerprint verification or face verification, multi-level security verification is realized, significantly enhancing the security of the transaction, reducing the problems of poor user experience and reduced efficiency caused by over-verification or unreasonable verification processes, and ensuring the efficiency and smoothness of the transaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 is the method flow chart of the present invention;

[0064] Figure 2 is the flow chart of obtaining the user risk assessment result of the present invention;

[0065] Figure 3 is the flow chart of obtaining the user type classification result of the present invention;

[0066] Figure 4 is the flow chart of evaluating the deviation of purchase data of the present invention;

[0067] Figure 5 is the flow chart of obtaining the order risk assessment result of the present invention;

[0068] Figure 6 is the flow chart of obtaining the time period security analysis result of the present invention;

[0069] Figure 7 is the flow chart of obtaining the order transaction verification result of the present invention. Detailed implementation manners

[0070] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0071] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0072] Please refer to Figure 1 , the present invention provides a technical solution: an optimization method for an e-commerce transaction link based on risk analysis, including the following steps:

[0073] S1: Based on the e-commerce transaction platform, perform human-machine verification on the user, extract the time and number of errors spent by the user during the human-machine verification process, analyze the deviation from the normal verification data, and combine the number of logins and the type of login device of the user to evaluate the user risk index and obtain the user risk assessment result;

[0074] S2: Based on the transaction records of the user on the e-commerce transaction platform, compare the type of transaction goods, average purchase amount and purchase times with the transaction information of the known user types to identify the user type and obtain the user type classification result;

[0075] S3: Based on the user type classification result and the user risk assessment result, according to the order information of the user on the e-commerce transaction platform currently, including the number of orders, order amount and type of goods, extract the purchase data of the same type of goods by the same type of users, including the average purchase quantity and purchase amount, evaluate the deviation of the purchase data, and combine the user risk index to evaluate the transaction risk level of the current order and obtain the order risk assessment result;

[0076] S4: Based on the transaction records of the e-commerce transaction platform, count the transaction data of the e-commerce transaction platform in each time period, including the average number of transactions, average total transaction amount and average number of transactions in each time period, and combine the standard deviation and abnormal transaction frequency of the transaction data in each time period to evaluate the required security level in the current time period and obtain the time period security analysis result;

[0077] S5: Based on the order risk assessment result and the time period security analysis result, according to the transaction risk level of the order and the security level required for the current time period, evaluate the security verification level required by the user. According to the preset security verification standard, adjust the number of steps and the type of verification content in the user's security verification process, including SMS verification, fingerprint verification, and face verification. After the security verification, implement the order transaction to obtain the order transaction verification result.

[0078] The user risk assessment result includes the login risk index and the login device mark. The user type classification result includes the user's purchase preference, transaction frequency, and purchase amount category. The order risk assessment result includes the risk deviation value, transaction security score, and potential risk assessment value. The time period security analysis result includes the security risk coefficient for each time period, the standard deviation of transaction data, and the high-risk time period mark. The order transaction verification result includes the adjusted security verification steps, the selection information of the verification type, and the adjusted information of the verification times.

[0079] Please refer to Figure 2 , the steps for obtaining the user risk assessment result are as follows:

[0080] S111: Based on the e-commerce trading platform, conduct a human-machine verification on the user, collect the data of the user during the human-machine verification process, including the user's verification time and the number of errors, to obtain the human-machine verification record:

[0081] The user conducts a human-machine verification through the e-commerce trading platform. During the verification process, the system collects the user's verification time and the number of errors data. The verification time refers to the time required for the user to complete one verification, in seconds. By calculating the comparison between the user's actual verification time and the average verification time of normal users, it is judged whether the user's performance is abnormal. The number of errors refers to the number of errors made by the user during the verification process, usually measured by the number of times of submitting incorrect verification codes or other mistakes. The number of errors of the user should not exceed the preset maximum value. Exceeding it indicates that the user's operation is abnormal. The collected verification time and the number of errors are important bases for the system to determine whether the user has risks. Establish a verification record database for users. Subsequently, the data of different users can be compared to evaluate their risk levels. Suppose that in one verification, the time for the user to complete the verification is 15 seconds, while the average verification time of normal users is 10 seconds, and the number of errors is 2 times, while the average number of errors of normal users is 1 time. It can be identified that the user's performance deviates from the normal level during verification and may have potential risks.

[0082] S112: Based on the human-machine verification record, collect the user's login device information, and detect whether the user's login device is a commonly used device and a known device. If it is a known commonly used device, mark it as 0, otherwise mark it as 1, to obtain the user login device identification score information;

[0083] Collect the user's device information for logging in and detect whether the device is a frequently used device and a known device. The key to the process lies in how to define "frequently used device" and "known device". A "frequently used device" usually refers to a device that the user has used for a long time and logs in frequently, such as the device bound when the user registers an account or the device that the user often logs in with. By checking the identification information of the device, such as the unique identifier (UID) of the device and the IP address of the device, combined with the user's historical login records, it is determined whether the device belongs to a frequently used device. If the device is a frequently used device, it will be marked as 0, indicating that the device is reliable; if the device is an infrequently used device, it will be marked as 1, indicating that the device may pose a risk. By comparing the user's historical login records and the identity information of the device, the trust level of the device is determined to ensure more accurate risk assessment. It may be necessary to compare the device identification and login history in the database. If the user often logs in through the same device, then the device is a frequently used device. For example, if a certain user A often logs in to his account through device 1, while device 2 has only occasional login records, the system will mark device 2 as 1 (unknown device) and give the corresponding risk identification.

[0084] S113: Based on the marked information of the user device status and the human-machine verification record, through the formula:

[0085]

[0086] Calculate the user risk index to obtain the user risk assessment result;

[0087] Among them, J user is the verification time of the user during the human-machine verification process, J avg is the average verification time of normal users, E user is the number of errors of the user during the human-machine verification process, E avg is the average number of errors of normal users, L R is the number of login attempts of the user, D type is the score of the user login device identification, α R is the weight coefficient of the deviation term, β R is the weight coefficient of the number of logins, γ R is the weight coefficient of the device type.

[0088] Formula:

[0089]

[0090] The advantage of the formula is that by comprehensively considering the user's verification performance, login times, and device identification, it can effectively evaluate the user's risk level, thus providing strong support for subsequent security decisions. Through weighted calculations of multiple factors, the formula makes the risk assessment more comprehensive and can accurately reflect the user's risk situation. By adjusting the weights of various parameters, different levels of attention can be given to different types of risks according to the actual situation, making the risk assessment more refined.

[0091] Detailed explanation of the formula and the derivation process of the formula calculation:

[0092] J user is the verification time of the user during the human-machine verification process, in seconds. J avg is the average verification time of normal users, in seconds. By comparing the difference between the user's verification time and the average verification time of normal users, the degree of abnormality of the user during the verification process can be obtained. For example, if the verification time of user A is 15 seconds and the average verification time of normal users is 10 seconds, the calculation result is

[0093] E user is the number of errors of the user during the human-machine verification process. E avg is the average number of errors of normal users. By comparing the number of errors of the user with the average number of errors of normal users, it can be reflected whether there are abnormalities in the user's operations. For example, if the number of errors of user A is 2 times and the average number of errors of normal users is 1 time, the calculation result is

[0094] L R is the number of login attempts of the user. Suppose user A has made 5 login attempts in the past week. D type is the identification score of the user's login device. Suppose user A is using a known device and the device identification score is 0.

[0095] α R 、β R 、γ R are the weight coefficients of each parameter respectively, and the coefficients need to be set according to the actual situation. For example, α R = 0.3, β R = 0.2, γ R = 0.5.

[0096] Substitute each parameter into the formula for calculation:

[0097] R = 0.3·(0.5 + 1) + 0.2·5 + 0.5·0 = 0.3·1.5 + 1 + 0 = 1.45;

[0098] The results show that the risk index of User A is 1.45. Based on the risk assessment results, it can be determined whether User A needs additional verification or other security measures. If this value exceeds the set risk threshold, investigation or follow-up measures may be required.

[0099] Please refer to Figure 3 , the steps to obtain the user type classification results are as follows:

[0100] S211: Based on the transaction records of the user on the e-commerce trading platform, extract the types of goods traded by the user, the average purchase amount, and the number of purchases to obtain the user's purchase characteristic data;

[0101] Extract relevant data from the user's transaction records on the e-commerce trading platform, including the types of goods traded, the average purchase amount, and the number of purchases, to construct the user's purchase characteristic data. The type of goods purchased refers to the category of goods purchased by the user, which can usually be identified according to the classification identifier or type label of the goods. The average purchase amount refers to the average amount of each transaction of the user within a specific time period, which is obtained by summing up all the purchase amounts of the user and then dividing by the number of purchases. The number of purchases is the number of transactions made by the user during this time period. For example, assume that User A has purchased 5 different types of goods in the past month, and the purchase amounts of each type of goods are 100 yuan, 150 yuan, 200 yuan, 250 yuan, and 300 yuan respectively. Then, the average purchase amount of User A is calculated as At the same time, the number of purchases of User A is 5. By obtaining the user's purchase characteristic data, the data will serve as the basis for subsequent risk assessment and user classification.

[0102] S212: Based on the user's purchase characteristic data, compare it with the transaction information of known user types through the formula:

[0103]

[0104] Calculate the matching degree of the user type;

[0105] where B is the matching degree of the user type, w i is the weight coefficient of the i-th feature, P user，i is the purchase data of the user on the i-th feature, P known，i is the purchase data of the known user type on the i-th feature, and n is the total number of features;

[0106] Formula:

[0107]

[0108] The advantage of the formula is that by comprehensively comparing the purchase characteristic data of users with the transaction information of known user types, it can accurately evaluate the matching degree of user types, thus better classifying users and further improving the accuracy of risk management. By introducing the weight coefficient w i assigning different influences to different characteristics can adjust the matching degree of specific characteristics according to the actual situation, making the calculation of the matching degree of user types more flexible and adaptable. This formula combines multi-dimensional data of user purchase behavior and helps to achieve personalized user classification and risk assessment.

[0109] Detailed explanation of the formula and the derivation process of formula calculation:

[0110] B is the matching degree of the user type, indicating the matching degree between the user and the known type. The larger the value, the higher the matching degree.

[0111] w i is the weight coefficient of the i-th characteristic, indicating the relative importance of this characteristic in the calculation of the matching degree. For example, if the commodity type has a greater impact on user behavior, a higher weight can be assigned to it.

[0112] P user，i is the purchase data of the user on the i-th characteristic (such as purchase amount or purchase frequency, etc.), which is obtained by extracting the corresponding characteristic data from the user's transaction records.

[0113] P known，i is the purchase data of the known user type on the i-th characteristic, usually from an existing user behavior data model or historical data.

[0114] n is the total number of characteristics, that is, the number of characteristics considered when calculating the matching degree. Suppose three characteristics (purchase amount, purchase frequency, purchase commodity type) are considered, then n = 3.

[0115] Suppose when calculating the matching degree between user A and a certain known user type, the purchase amount of user A is 200 yuan, the purchase amount of the known user type is 150 yuan, the purchase frequency is 5 times, the purchase frequency of the known user type is 4 times, and there are also differences in the commodity type characteristics between the two. Suppose the corresponding weight coefficients are w 1 = 0.5, w 2 = 0.3, w 3 = 0.2, and the matching degree between user A and the known user type in terms of commodity type is 0.1. Then according to the formula, the calculation process of the matching degree is as follows:

[0116]

[0117] The result shows that the matching degree of User A with known user types is 0.2617. Based on this matching degree, the system can determine the type of User A, compare it with other types, and select the most matching user type as its classification result.

[0118] S213: Based on the matching degree of user types, by comparing the magnitudes of the matching degrees of each user type, select the user type with the maximum match as the matching type to obtain the user type classification result;

[0119] Based on the matching degree of user types, by comparing the matching degrees of each user type, select the type with the maximum matching degree with the user as the user type classification result. The process will sort the calculated matching degrees of the users and make a judgment by selecting the type with the highest matching degree. For example, if the matching degree of a user with a certain known type is relatively high (such as 0.8), then the user is classified as this type. The core of the operation lies in comparing the matching degrees of different user types and selecting the most suitable type, which can ensure that users are correctly classified and provide accurate data support for subsequent personalized recommendations and marketing strategies.

[0120] Please refer to Figure 4 , the steps to evaluate the purchase data deviation are as follows:

[0121] S311: Collect the order information of the user on the e-commerce trading platform currently, including the order quantity, order amount, and goods type, to obtain the user order data;

[0122] Collect the order information of the user on the e-commerce trading platform, obtain data including the order quantity, order amount, and goods type, etc., and construct the user's order data. The order quantity refers to the number of orders placed by the user within a certain time range, the order amount refers to the total amount paid by the user in all orders, and the goods type refers to the categories of goods purchased by the user in the orders. Through these data, the purchase behavior and order characteristics of the user can be reflected, providing a basis for subsequent risk assessment and user behavior analysis. For example, User A placed 10 orders in the past month with a total amount of 2000 yuan, including several different types of goods such as electronic products and household items. By counting these order data, it can be obtained that the order quantity of User A is 10, the order amount is 2000 yuan, and the goods type includes two categories: electronic products and household items. This data will be used as the basis for the subsequent evaluation of the user.

[0123] S312: Based on the user order data and the user type classification result, extract the purchase data of similar users, including the average purchase quantity and average purchase amount of similar goods, to obtain the purchase data of similar users;

[0124] By extracting the purchase data of similar users for comparison, the average purchase quantity and average purchase amount of similar goods are obtained. According to the classification result of user types, users are divided into different types, the purchase data of similar users are extracted, the similar users are clustered, and users with similar purchase behaviors and attributes are grouped into the same category. The average purchase quantity and average purchase amount of similar users for specific goods are calculated. For example, assume that the purchase data of user A is 10 orders with a total amount of 2,000 yuan, and the types of purchased goods include electronic products, household goods, etc. By retrieving other users with similar purchase behaviors to user A, the average purchase quantity and purchase amount of users in categories such as electronic products and household goods are calculated to obtain the purchase data of similar users. For example, the average purchase quantity of similar users for electronic products is 5, and the average purchase amount is 1,000 yuan; the average purchase quantity of similar users for household goods is 4, and the average purchase amount is 800 yuan. The purchase characteristics of the user type to which user A belongs can be obtained, thus providing data support for subsequent deviation evaluation.

[0125] S313: Based on the user order data and the purchase data of similar users, through the formula:

[0126]

[0127] Calculate the purchase data deviation index. According to the magnitude of the deviation index, evaluate the degree of deviation to obtain the purchase data deviation evaluation result;

[0128] where M is the purchase data deviation index, Q user is the purchase quantity of the user in the current order, Q avg is the average purchase quantity of similar users for this type of goods, A user is the purchase amount of the current user in the current order, A avg is the average purchase amount of similar users for this type of goods, μ M is the index coefficient;

[0129] Formula:

[0130]

[0131] The advantage of this formula is that by calculating the deviation index of the purchase data, it can effectively measure whether the user's purchase behavior is abnormal, and evaluate the risk level of the user through this deviation index. This formula not only considers the deviation of the user's purchase quantity but also the deviation of the purchase amount, and at the same time introduces the index coefficient μ M , making the calculation of the deviation more precise and assigning higher weights to data with larger deviations. This method can help the system more accurately identify users who deviate from the normal purchase behavior.

[0132] Detailed explanation of the formula and the formula calculation derivation process:

[0133] M is the purchase data deviation index, which represents the deviation degree between the user's purchase behavior and that of similar users.

[0134] Q user is the purchase quantity of the current user in the order, and Q avg is the average purchase quantity of similar users for this type of goods. By calculating the deviation between the user's purchase quantity and the average purchase quantity of similar users, the abnormality degree of the user's purchase behavior is measured.

[0135] A user is the purchase amount of the current user in the order, and A avg is the average purchase amount of similar users for this type of goods. By calculating the deviation between the user's purchase amount and the average purchase amount of similar users, the user's purchase behavior is further evaluated.

[0136] μ M is the exponential coefficient, which is used to weight the deviation and is usually set as a constant greater than 1 to strengthen the calculation results with larger deviations.

[0137] Suppose the purchase quantity of user A for electronic products is 10, and the purchase amount is 2000 yuan. The average purchase quantity of similar users is 8, and the average purchase amount is 1500 yuan, and the exponential coefficient is μ M = 2. Then the calculation process of the purchase data deviation index is as follows:

[0138]

[0139] The result shows that the purchase data deviation index of user A is 0.1736. According to this deviation index, it can be judged that there is a certain degree of deviation in the purchase behavior of user A compared with that of similar users. This deviation index can be used to evaluate the risk level of the user. If the deviation index is too high, it indicates that the user's purchase behavior is relatively abnormal and may require review or additional measures.

[0140] Please refer to Figure 5 for the steps to obtain the order risk assessment result:

[0141] S321: Based on the user risk assessment result and the purchase data deviation assessment result, according to the order risk management standard, extract the preset minimum transaction risk level and maximum transaction risk level to obtain the transaction risk association data;

[0142] Based on the user's risk assessment results and the purchase data deviation assessment results, combined with the order risk management standards, the preset minimum transaction risk level and the maximum transaction risk level are extracted. The preset minimum transaction risk level and the maximum transaction risk level are usually obtained through statistical analysis of historical data, and the thresholds for specific transaction risk levels are calibrated. The data is used as the basis for subsequent risk assessments. Through the preset levels, a preliminary risk level range can be assigned to each user or each order, providing a basis for risk management operations. For example, based on historical data analysis, the minimum transaction risk level is set to 1 and the maximum transaction risk level is set to 10 to evaluate the risk of orders. The order will be classified into this level range according to the specific risk assessment results of the order.

[0143] S322: Based on the transaction risk correlation data, through the formula:

[0144]

[0145] Calculate the transaction risk level of the current order to obtain the order risk assessment result;

[0146] Among them, L order is the transaction risk level of the current order, L min is the preset minimum transaction risk level, L max is the preset maximum transaction risk level, M is the purchase data deviation index, R is the user risk index, M max is the preset maximum purchase data deviation value, R max is the preset maximum user risk index, α L and β L are weight coefficients.

[0147] Formula:

[0148]

[0149] The advantage of the formula is that by comprehensively considering the purchase data deviation index and the risk index of the user, the formula can accurately evaluate the transaction risk level of each order. By introducing the weighted combination of the purchase data deviation index M and the user risk index R, it can effectively reflect the deviation and risk of user behavior, so as to assign a suitable risk level to each order. The formula combines the two risk factors, and through the maximum value function and the rounding-up operation, it ensures that the risk level of the order is within a reasonable range and provides accurate data support for subsequent risk control measures.

[0150] Detailed explanation of the formula and the formula calculation derivation process:

[0151] L order is the transaction risk level of the current order, indicating the risk degree of the order. The higher the risk level, the greater the risk of the order.

[0152] L min is the preset minimum transaction risk level, usually 1, indicating the lowest risk level.

[0153] L max is the preset maximum transaction risk level, usually 10, indicating the highest risk level.

[0154] M is the purchase data deviation index, representing the degree of deviation between the purchase data of the current order and that of similar users. The larger the value, the more abnormal the purchase behavior of the order.

[0155] M max is the preset maximum purchase data deviation value, representing the maximum allowed deviation, usually obtained through historical data analysis.

[0156] R is the user risk index, representing the risk assessment value of the user, calculated based on multiple factors such as their historical transaction behavior and account security.

[0157] R max is the preset maximum user risk index, usually 5, indicating the highest user risk level.

[0158] α L and β L are weight coefficients, respectively representing the influence degrees of the purchase data deviation index and the user risk index on the order risk level. The coefficients can be adjusted according to the actual situation, usually determined according to different industry requirements, risk preferences, etc.

[0159] Suppose: L min = 1, L max = 10, M = 0.5, M max = 1, R = 2, R max = 5, α L = 0.6, β L = 0.4.

[0160] Substitute the above data into the formula for calculation:

[0161]

[0162] The result shows that the transaction risk level of the current order is 5. Based on this risk level, it will be determined whether to take additional security measures, such as verifying the user's identity, restricting the order payment method, etc.

[0163] Please refer to Figure 6 , the steps to obtain the time period security analysis result are:

[0164] S411: Based on the transaction records of the e-commerce trading platform, statistically analyze the transaction data of the e-commerce trading platform in each time period, including the average number of transactions, the average total transaction amount, and the average number of transaction users in each time period, calculate the standard deviation of the transaction data in each time period, and statistically analyze the abnormal transaction frequency in each time period to obtain the statistical information of the time period transaction data;

[0165] Statistically analyze the transaction data in each time period, including indicators such as the average number of transactions, the average total transaction amount, and the average number of transaction users in each time period. The number of transactions refers to the number of successful transactions by users within each time period. The transaction amount refers to the total amount of all orders within that time period. The number of transaction users refers to the number of independent users who have conducted at least one transaction within that time period. The standard deviation is an indicator to measure the degree of dispersion of the transaction data distribution. The larger the standard deviation, the greater the volatility of the transaction data within that time period. The abnormal transaction frequency refers to the proportion of abnormal transactions (such as high-value transactions or frequent transactions, etc.) within that time period. By statistically analyzing the indicators, the platform can better analyze the characteristics of transaction activities in each time period. The statistical data helps to analyze the transaction characteristics of that time period and provides support for subsequent decision-making.

[0166] S412: Based on the statistical information of the time period transaction data, through the formula:

[0167]

[0168] Calculate the data deviation coefficient of the current time period;

[0169] where S is the data deviation coefficient of the current time period, T cu is the average number of transactions in the current time period, T avg is the average number of transactions in all time periods, A cu is the average transaction amount in the current time period, A avg is the average transaction amount in all time periods, N cu is the average number of transaction users in the current time period, N avg is the average number of transaction users in all time periods, STD cu is the standard deviation of the transaction data in the current time period, STD avg is the standard deviation of the transaction data in all time periods, F cu is the average abnormal transaction frequency in the current time period, F avg is the average abnormal transaction frequency in all time periods, α W β W γ W δ W ∈ W are weight coefficients;

[0170] S413: Based on the data deviation coefficient of the current time period, through the formula:

[0171]

[0172] Calculate the required safety level for the current period to obtain the period safety analysis result;

[0173] where L required is the required safety level for the current period, L base is the reference safety level value, and S is the data deviation coefficient for the current period;

[0174] Formula:

[0175]

[0176] The advantage of the formula is that by comprehensively considering multiple important indicators, the formula can accurately calculate the data deviation coefficient for the current period, thereby measuring the degree of deviation of this period from the overall trend. By comparing the number of transactions, transaction amount, number of traders, standard deviation, and abnormal transaction frequency of the period with the average value of all periods and weighting according to the weight coefficients of each indicator, the differences in the transaction data of the current period from other periods can be comprehensively evaluated. When evaluating the transaction data of a period, the formula does not only consider a single deviation factor, but takes into account data from multiple dimensions, ensuring the comprehensiveness and accuracy of the deviation coefficient.

[0177] Detailed explanation of the formula and the formula calculation derivation process:

[0178] S is the data deviation coefficient for the current period, indicating the degree of deviation of the transaction data in the current period from the overall trend.

[0179] T cu is the average number of transactions in the current period, T avg is the average number of transactions in all periods, and the ratio of the two represents the comparison of the transaction activity in the current period with the overall situation.

[0180] A cu is the average transaction amount in the current period, A avg is the average transaction amount in all periods, and the ratio of the two reflects the difference in the transaction amount level in the current period from other periods.

[0181] N cu is the average number of traders in the current period, N avg is the average number of traders in all periods, and the ratio of the two reflects the user activity in the current period.

[0182] STD cu is the standard deviation of the transaction data in the current period, STD acg is the standard deviation of the transaction data in all periods, and the ratio of the two reflects the comparison of the volatility of the transaction data in the current period with the overall situation.

[0183] F cu is the average abnormal trading frequency for the current period, and F avg is the average abnormal trading frequency for all periods. The ratio of the two reflects the frequency of abnormal trading in the current period.

[0184] α W , β W , γ W , δ W , ∈ W are the weight coefficients of each indicator, indicating the influence degree of different indicators on the final result.

[0185] Let: T cu = 120 (average number of transactions in the current period), T avg = 100 (average number of transactions in all periods), A cu = 550 (average transaction amount in the current period), A avg = 500 (average transaction amount in all periods), N cu = 60 (average number of transaction participants in the current period), N avg = 50 (average number of transaction participants in all periods), STD cu = 7000 (standard deviation of transaction amount in the current period), STD avg = 5000 (standard deviation of transaction amount in all periods), F cu = 0.07 (abnormal trading frequency in the current period), F avg = 0.05 (average abnormal trading frequency in all periods), weight coefficient: α W = 0.2, β W = 0.2, δ W = 0.2, δ W = 0.2, ∈ W = 0.2.

[0186] Substitute the above data into the formula for calculation:

[0187]

[0188] The result shows that the data deviation coefficient for the current period is 1.26, indicating that the transaction data for this period deviates to a certain extent from the data for all periods.

[0189] Please refer to Figure 7 , and the steps to obtain the order transaction verification result are as follows:

[0190] S511: Based on the order risk assessment result and the period security analysis result, according to the transaction risk level of the order and the required security level for the current period, through the formula:

[0191]

[0192] Calculate the security verification level required by the user;

[0193] where L ver is the security verification level required by the user, L Z is the baseline security verification level, L order is the transaction risk level of the current order, L oavg is the average transaction risk level of all orders, L required is the required security level for the current period, L ravg is the average required security level for all periods, α v and β v are weight coefficients;

[0194] Formula:

[0195]

[0196] The advantage of the formula is that by comprehensively considering the transaction risk level of the order and the security level of the current period, it can accurately calculate the security verification level required by the user. This method takes into account two factors: the risk of the order itself and the security requirements of the period, thus ensuring the accurate setting of the verification level. At the same time, the weight coefficients α v and β v are introduced, enabling the influence of different risk factors to be adjusted according to the actual situation, thereby enhancing the flexibility and accuracy of security assessment.

[0197] Detailed explanation of the formula and the derivation process of formula calculation:

[0198] L ver is the security verification level required by the user, representing the security verification level calculated based on the order risk and the security requirements of the period.

[0199] L Z is the baseline security verification level, usually a preset standard value used to calibrate the basic requirements of security verification.

[0200] L order is the transaction risk level of the current order, reflecting the risk degree of the order itself.

[0201] L oavg is the average transaction risk level of all orders, used to measure the risk of the current order relative to the overall risk level.

[0202] L required is the required security level for the current period, evaluated based on the transaction activities and abnormal situations of the period.

[0203] L ravg is the average required security level for all time periods, used to compare the security requirements of the current time period.

[0204] α v and β v are weight coefficients, respectively used to adjust the influence of order risk and time period security level on the user's security verification level.

[0205] Let: L Z = 5 (benchmark security verification level), L order = 7 (transaction risk level of the current order), L oavg = 5 (average transaction risk level of all orders), L required = 8 (required security level for the current time period), L ravg = 6 (average required security level for all time periods), α v = 0.6, β v = 0.4.

[0206] Substitute the above data into the formula for calculation:

[0207]

[0208] The result shows that the required security verification level for the user is 7. According to the calculation, perform the corresponding security verification steps to ensure the security of the order.

[0209] S512: Based on the required security verification level of the user, according to the preset security verification standard, adjust the number of steps and the type of verification content in the user's security verification process, including SMS verification, fingerprint verification, and face verification. After the security verification, implement the order transaction to obtain the order transaction verification result;

[0210] According to the preset security verification standard, adjust the number of steps and the type of verification content in the user's verification process. The number of verification steps and the type of content include SMS verification, fingerprint verification, and face verification, etc. The setting of the steps will be adjusted according to the current security verification level to ensure that the verification process is neither cumbersome nor too simple. In actual implementation, the required verification steps will be determined by analyzing the user's security verification level. For example, when the user's security verification level is high, the verification steps will be increased, such as requiring both fingerprint verification and face verification at the same time to improve security. When the security verification level is low, only SMS verification may be required. For example, if the calculated security verification level of user A is 7, in this case, user A will be required to perform face verification and SMS verification to confirm their identity to ensure the security of the transaction. After the user completes these verifications, implement the order transaction to generate the order transaction verification result.

[0211] An e-commerce transaction link optimization system based on risk analysis, which is used to execute the above-mentioned e-commerce transaction link optimization method based on risk analysis. The system includes:

[0212] The user risk analysis module, based on the e-commerce transaction platform, conducts human-machine verification on users, extracts the time spent and the number of errors made by users during the human-machine verification process, analyzes the deviation from the normal verification data, and combines the number of login times and the type of login device of the users to evaluate the user risk index and obtain the user risk assessment result.

[0213] The user type identification module, based on the transaction records of users on the e-commerce transaction platform, compares the types of traded goods, average purchase amount, and purchase times with the transaction information of known user types to identify the user type and obtain the user type classification result.

[0214] The order risk analysis module, based on the user type classification result and the user risk assessment result, extracts the purchase data of similar users for the same type of goods according to the current order information of the user on the e-commerce transaction platform, evaluates the deviation of the purchase data, and combines the user risk index to evaluate the transaction risk level of the current order and obtain the order risk assessment result.

[0215] The time period impact analysis module, based on the transaction records of the e-commerce transaction platform, counts the transaction data of the e-commerce transaction platform in each time period, and combines the standard deviation and abnormal transaction frequency of the transaction data in each time period to evaluate the required security level of the current time period and obtain the time period security analysis result.

[0216] The transaction verification implementation module, based on the order risk assessment result and the time period security analysis result, evaluates the required security verification level of the user according to the transaction risk level of the order and the required security level of the current time period, adjusts the number of steps and the type of verification content in the user's security verification process, and implements the order transaction after the security verification to obtain the order transaction verification result.

[0217] The above is only the preferred embodiment of the present invention, and it does not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An e-commerce transaction link optimization method based on risk analysis, characterized in that: The following steps are involved: S1: Based on the e-commerce transaction platform, perform human-machine verification on users, extract the time and number of errors spent by users in the human-machine verification process, and evaluate the user risk index based on the number of user logins and login device type to obtain the user risk assessment result; S2: Based on the transaction records of users on the e-commerce trading platform, the transaction information of known user types is compared to identify the user type and obtain the user type classification result; S3: Based on the user type classification result and the user risk assessment result, according to the user's current order information on the e-commerce trading platform, extract the purchase data of the same type of users for the same type of goods, evaluate the purchase data deviation, and combine the user risk index to evaluate the transaction risk level of the current order to obtain the order risk assessment result; S4: Based on the transaction records of the e-commerce trading platform, the transaction data of the e-commerce trading platform in each time period is counted, and the standard deviation of the transaction data in each time period and the abnormal transaction frequency are combined to evaluate the security level required for the current time period and obtain the security analysis results of the time period; S5: Based on the order risk assessment results and the time period security analysis results, according to the transaction risk level of the order and the security level required for the current time period, evaluate the security verification level required by the user, adjust the number of steps and verification content type in the user's security verification process, and after the security verification, implement the order transaction to obtain the order transaction verification result.

2. The method for optimizing e-commerce transaction links based on risk analysis according to claim 1, characterized in that: The steps for obtaining the user risk assessment result are: S111: Based on the e-commerce transaction platform, perform human-machine verification on the user, collect the user's data during the human-machine verification process, including the user's verification time and number of errors, and obtain the human-machine verification record: S112: Based on the human-machine verification record, collect user login device information, detect whether the user login device is a common device and a known device, if it is a known common device, mark it as 0, otherwise mark it as 1, and obtain user login device identification score information; S113: Based on the user device status mark information and the human-machine verification record, the formula: Calculate the user risk index and obtain the user risk assessment result; Among them, J user is the verification time of the user during the human-machine verification process, J avg is the average verification time of normal users, E user is the number of errors made by the user during human-machine verification, E avg is the average number of errors for normal users, L R is the number of login attempts of the user, D type The user login device identification score, α R is the weight coefficient of the bias term, β R is the weight coefficient of the number of logins, γ R is the weight coefficient of the device type.

3. The method for optimizing e-commerce transaction links based on risk analysis according to claim 1, characterized in that: The steps for obtaining the user type classification result are: S211: Based on the transaction records of the user on the e-commerce transaction platform, extract the user's transaction goods type, average purchase amount and purchase frequency to obtain the user's purchase feature data; S212: Based on the user purchase feature data, compare with the transaction information of known user types, through the formula: Calculate the matching degree of user type; Among them, B is the matching degree of user type, w i is the weight coefficient of the i-th feature, P user，i is the user’s purchase data on the i-th feature, P known，i is the purchase data of the known user type on the i-th feature, and n is the total number of features; S213: Based on the matching degree of the user types, by comparing the matching degree of each user type, the user type with the greatest matching is selected as the matching type, and a user type classification result is obtained.

4. The method for optimizing e-commerce transaction links based on risk analysis according to claim 1, characterized in that: The steps for evaluating purchase data deviation are: S311: Collect the user's current order information on the e-commerce transaction platform, including order quantity, order amount and product type, to obtain user order data; S312: Based on the user order data and the user type classification result, extract the purchase data of the same type of users, including the average purchase quantity and average purchase amount of the same type of goods, to obtain the purchase data of the same type of users; S313: Based on the user order data and the purchase data of similar users, the formula: Calculate the purchase data deviation index, evaluate the degree of deviation according to the size of the deviation index, and obtain the purchase data deviation evaluation result; Among them, M is the purchase data deviation index, Q user is the purchase quantity of the user in the current order, Q avg is the average purchase quantity of the same type of goods by the same type of users, A user A is the purchase amount of the current user in the current order, avg is the average purchase amount of the same type of users on the same type of goods, μ M is the exponential coefficient.

5. The method for optimizing e-commerce transaction links based on risk analysis according to claim 4 is characterized in that: The steps for obtaining the order risk assessment results are as follows: S321: Based on the user risk assessment result and the purchase data deviation assessment result, according to the order risk management standard, extract the preset minimum transaction risk level and the maximum transaction risk level to obtain transaction risk associated data; S322: Based on the transaction risk association data, by formula: Calculate the transaction risk level of the current order and obtain the order risk assessment result; Among them, L order is the transaction risk level of the current order, L min is the preset minimum transaction risk level, L max is the preset highest transaction risk level, M is the purchase data deviation index, R is the user risk index, and M max is the preset maximum purchase data deviation value, R max is the preset maximum user risk index, α L and β L is the weight coefficient.

6. The method for optimizing e-commerce transaction links based on risk analysis according to claim 1, characterized in that: The steps for obtaining the safety analysis results of the time period are: S411: Based on the transaction records of the e-commerce transaction platform, the transaction data of the e-commerce transaction platform in each time period is counted, including the average number of transactions, the average total transaction amount and the average number of transaction persons in each time period, the standard deviation of the transaction data in each time period is calculated, and the abnormal transaction frequency in each time period is counted to obtain the transaction data statistical information of the time period; S412: Based on the statistical information of the transaction data during the period, the formula: Calculate the data deviation coefficient for the current period; Among them, S is the data deviation coefficient of the current period, T cu is the average number of transactions in the current period, T avg is the average number of transactions in all periods, A cu is the average transaction amount in the current period, A avg is the average transaction amount in all periods, N cu is the average number of traders in the current period, N avg is the average number of traders in all periods, STD cu The standard deviation of the transaction data in the current period, STD avg is the standard deviation of transaction data for all periods, F cu is the average abnormal transaction frequency in the current period, F avg is the average abnormal transaction frequency in all periods, α W , β W , γ W ,δ W ,∈ W is the weight coefficient; S413: Based on the data deviation coefficient of the current period, the formula: Calculate the security level required for the current period and obtain the security analysis result of the period; Among them, L required is the security level required for the current period, L base is the benchmark safety level value, and S is the data deviation coefficient of the current period.

7. The method for optimizing e-commerce transaction links based on risk analysis according to claim 1, characterized in that: The steps for obtaining the order transaction verification result are: S511: Based on the order risk assessment result and the period security analysis result, according to the transaction risk level of the order and the security level required for the current period, the formula is: Calculate the security verification level required by the user; Among them, L ver is the security verification level required by the user, L Z is the baseline safety verification level, L order is the transaction risk level of the current order, L oavg is the average transaction risk level of all orders, L required is the required security level for the current period, L ravg is the average required security level for all time periods, α v and β v is the weight coefficient; S512: Based on the security verification level required by the user, in accordance with the preset security verification standards, adjust the number of steps and verification content types in the user's security verification process, including SMS verification, fingerprint verification and face verification. After the security verification, implement the order transaction and obtain the order transaction verification result.

8. An e-commerce transaction link optimization system based on risk analysis, characterized in that: According to the method for optimizing e-commerce transaction links based on risk analysis according to any one of claims 1 to 7, the system comprises: The user risk analysis module performs human-machine verification on the user based on the e-commerce transaction platform, extracts the time and number of errors spent by the user in the human-machine verification process, and evaluates the user risk index based on the number of logins and login device type of the user to obtain the user risk assessment result; The user type identification module identifies the user type based on the user's transaction records on the e-commerce transaction platform, and compares the transaction information of known user types according to the transaction goods type, average purchase amount and purchase frequency, and obtains the user type classification result; The order risk analysis module extracts the purchase data of the same type of goods by the same type of users based on the user type classification result and the user risk assessment result according to the user's current order information on the e-commerce trading platform, evaluates the purchase data deviation, and evaluates the transaction risk level of the current order in combination with the user risk index to obtain the order risk assessment result; The time period impact analysis module is based on the transaction records of the e-commerce trading platform, and counts the transaction data of the e-commerce trading platform in each time period. It combines the standard deviation of the transaction data in each time period with the frequency of abnormal transactions, evaluates the security level required for the current time period, and obtains the security analysis results of the time period. The transaction verification implementation module is based on the order risk assessment results and the time period security analysis results. According to the transaction risk level of the order and the security level required for the current time period, it evaluates the security verification level required by the user, adjusts the number of steps and verification content type in the user's security verification process, and implements the order transaction after security verification to obtain the order transaction verification result.

Images