E-commerce digitalized intelligent risk control system and method

Abstract

The invention relates to the technical field of intelligent risk control, and discloses an intelligent risk control system and method for e-commerce digitization, and the system comprises a data acquisition module, a feature extraction module, an intelligent early warning module, and a risk assessment module. Abnormal conditions can be accurately identified; the intelligent early warning and risk assessment module performs double control, the former quickly grades and judges the risk based on a threshold value, and the latter calculates a comprehensive risk value through normalization processing and weight distribution to ensure that the risk judgment is accurate and reliable; graded early warning signals are triggered according to different risk grades, and are accurately pushed to related responsible parties, so that cross-department collaborative prevention and control is realized, and the response efficiency is improved. Meanwhile, the system continuously optimizes the evaluation model based on historical data, so that the risk control strategy is more suitable for the business reality, the risk resistance of the e-commerce platform is enhanced, and the transaction security and the healthy development of the industry are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent risk control, and particularly relates to an intelligent risk control system and method for e-commerce digitization. Background Art

[0002] With the rapid development of Internet technology, the e-commerce industry has shown an explosive growth trend, with the transaction scale continuously expanding and the participating entities increasing day by day. However, while the e-commerce business is booming, various transaction risks have also emerged, such as frequent problems like merchant fraud, false transactions, payment anomalies, and logistics delays, which have seriously affected the healthy development of e-commerce platforms and the legitimate rights and interests of users. Most traditional e-commerce risk control systems rely on single-dimensional data or simple rule engines for risk judgment. For example, they only identify abnormal orders based on the transaction data of the e-commerce platform itself, or determine risks by setting fixed thresholds. This approach has obvious limitations. Due to the single data source, it cannot comprehensively reflect the real risk situation in the transaction process, easily misses potential risks, and leads to misjudgment or missed judgment of risks. At the same time, traditional rule engines lack adaptability to complex business scenarios and dynamic changes, and it is difficult to adjust risk control strategies in a timely manner with the development of e-commerce business and the evolution of risk forms, unable to meet the accuracy and real-time requirements of the e-commerce industry for risk prevention and control. In addition, among the multiple parties involved in e-commerce transactions, the data between e-commerce platforms, merchants, payment institutions, and logistics enterprises are often independent of each other, lacking an effective integration and sharing mechanism. This makes it impossible for all parties to obtain comprehensive and complete information when conducting risk assessments, difficult to form a comprehensive judgment on transaction risks, and reduces the efficiency and effectiveness of risk control. Therefore, there is an urgent need for an e-commerce digital intelligent risk control system that can integrate multi-party data resources, use advanced algorithms and models to achieve accurate and dynamic risk assessment and early warning to cope with the increasingly complex and changeable e-commerce transaction risks. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent risk control system and method for e-commerce digitization, which solves the technical problems proposed in the background art.

[0004] The purpose of the present invention can be achieved through the following technical solutions: An intelligent risk control system for e-commerce digitization, comprising: A data acquisition module, configured to acquire various key data of goods in real time from four parties: e-commerce platforms, merchants, payment institutions, and logistics; A feature extraction module, configured to extract combined feature variables for depicting the risk situation based on various key data, and the combined feature variables include a merchant transaction fluctuation index, a logistics timeliness deviation rate, a payment anomaly coefficient, and an operating index deviation degree; An intelligent early warning module for classifying the levels of e-commerce transaction risks based on combined feature variables; A risk assessment module for comprehensively evaluating the comprehensive risk level of a transaction according to the extracted combined feature variables.

[0005] As a further solution of the present invention: various key data include: Transaction data of the e-commerce platform, specifically covering: order ID, user ID, order amount, order time stamp, user's IP address for placing the order; Business operation data of the merchant, specifically covering: merchant ID, product category code, product price, inventory quantity, number and amount of historical transactions; Payment data of the payment institution, specifically covering: payment order ID, payment channel code, payment amount, payment time, payment response status code; Among them, the payment response status code is a digital code returned by the payment gateway, bank or third-party payment platform after processing a transaction request, used to clearly indicate the processing status and result of the payment transaction; the processing status and result of the payment transaction specifically refer to the situations of identifying successful payment, failed payment or abnormal payment response; Logistics data of the logistics, specifically covering: waybill ID, pick-up time, transportation node information, receipt time, logistics service provider code.

[0006] As a further solution of the present invention: the extraction method of the combined feature variables is as follows: Step A1, merchant transaction fluctuation index: Based on the time sequence relationship between the order time stamp in the transaction data and the current system time, perform two-dimensional order quantity statistics, specifically as follows: Count the order quantity within a preset standard observation window pushed forward from the current time node and mark it as D1; Count the average order quantity of each preset standard observation window within a preset specified period in the past and mark it as DP; Through: ; Calculate the merchant transaction fluctuation index M; Step A2, logistics timeliness deviation rate: Select a waybill ID, based on the time difference between the pick-up time and the receipt time in the logistics data corresponding to the waybill ID, determine the actual delivery time of the waybill ID and mark it as P1; Based on the user's IP address for placing the order corresponding to the product of the waybill ID, obtain multiple historical waybill IDs in the logistics data corresponding to the same user's IP address for placing the order, and obtain the time difference between the pick-up time and the receipt time therefrom, calculate the average delivery time of multiple historical waybills and mark it as PP; Through: ; Calculate the logistics timeliness deviation rate L; Step A3, Payment anomaly coefficient: Based on the payment time in the payment data, obtain all payment order IDs within a specified observation period, and record the number of all payment order IDs as the total number of payments Z0; At the same time, based on the payment response status code, count the payment order IDs with unsuccessful payments, that is, the payment order IDs with payment failure or payment anomaly response situations; and record all payment order IDs with unsuccessful payments as the number of payment anomalies Z1; By: ; Calculate the payment anomaly coefficient Y; Step A4, Deviation degree of business indicators: Based on the historical number of transactions in the business data, combined with the time sequence relationship of the current system time, count the number of transactions within a preset standard period pushed forward from the current time node, and mark it as B0; Count the number of transactions in each preset standard period within the past preset observation period, and mark it as Bt, where t = 1, 2,..., k, and k represents the number of standard periods within the specified period; By: ; Calculate the deviation degree of business indicators H.

[0007] As a further solution of the present invention: The level classification method of the intelligent early warning module is as follows: When M ∈ [Ma, Mb), and at the same time L < La, Y < Ya, and H < Ha all hold, it is determined that the e-commerce transaction risk is a first-level risk; When any one of the conditions M ∈ [Mb, Mc), L ∈ [La, Lb), Y ∈ [Ya, Yb), H ∈ [Ha, Hb) is satisfied, it is determined that the e-commerce transaction risk is a second-level risk; When at least one of the conditions M ≥ Mc, L ≥ Lb, Y ≥ Yb, H ≥ Hb holds, it is determined that the e-commerce transaction risk is a third-level risk; Where: Ma, Mb, Mc are thresholds preset according to the merchant transaction fluctuation index; La, Lb are thresholds preset according to the logistics timeliness deviation rate; Ya, Yb are thresholds preset according to the payment anomaly coefficient; Ha, Hb are thresholds preset according to the deviation degree of business indicators.

[0008] As a further solution of the present invention: The evaluation method of the comprehensive risk level is as follows: Step B1, Joint feature variable normalization processing: In historical data, extract multiple sets of various key data similar to the commodity. Subsequently, through the feature refinement module, extract the corresponding combined feature variables from the multiple sets of various key data in the historical data. Then, extract the minimum and maximum values of each combined feature variable and form a combined feature variable interval. The combined feature variable interval covers the merchant transaction fluctuation interval [M min , M max , the logistics timeliness deviation rate interval [L min , L max , the payment anomaly interval [Y min , Y max , and the business indicator deviation degree interval [H min , H max ; By: ; Calculate the normalized merchant transaction fluctuation index M1, logistics timeliness deviation rate L1, payment anomaly coefficient Y1, and business indicator deviation degree H1; Step B2: Extraction of weights for each combined feature variable: Extract the corresponding weight coefficients pre-assigned based on the merchant transaction fluctuation index, logistics timeliness deviation rate, payment anomaly coefficient, and business indicator deviation degree, and mark them as w1, w2, w3, and w4 in sequence; Among them, w1 + w2 + w3 + w4 = 1; Step B3: Calculation of the comprehensive risk value: By: ; Calculate the comprehensive risk value R; Step B4: Evaluation of the comprehensive risk level: According to the comprehensive risk value R, divide the risk level into three levels: low risk, medium risk, and high risk; The specific division criteria are as follows: When Ra ≤ R < Rb, it is determined that the comprehensive risk level is the low-risk level; When Rb ≤ R < Rc, it is determined that the comprehensive risk level is the medium-risk level; When Rc ≤ R ≤ Rd, it is determined that the comprehensive risk level is the high-risk level; Among them, Ra, Rb, Rc, and Rd are preset comprehensive risk thresholds, and Ra < Rb < Rc < Rd.

[0009] As a further solution of the present invention: The intelligent warning module and the risk assessment module also trigger warning signals of different levels based on the results obtained respectively; When the e-commerce transaction risk is determined to be a first-level risk or the comprehensive risk level is determined to be a low risk level, a first-level risk signal is generated, and the first-level risk signal is used to send risk warning information to the e-commerce platform operators; When the e-commerce transaction risk is determined to be a second-level risk or the comprehensive risk level is determined to be a medium risk level, a second-level risk signal is generated, and the second-level risk signal is used to send early warning notices to the e-commerce platform operators and merchants; When the e-commerce transaction risk is determined to be a third-level risk or the comprehensive risk level is determined to be a high risk level, a third-level risk signal is generated, and the third-level risk signal is used to send urgent early warnings to the e-commerce platform operators, the risk control departments of payment institutions, and the risk monitoring departments of logistics companies.

[0010] An intelligent risk control method for e-commerce digitization, which is implemented through an intelligent risk control system for e-commerce digitization. The method includes the following steps: Obtain data: Real-time obtain various key data of goods; Extract features: Based on various key data, extract joint feature variables; Level assessment: Based on the joint feature variables, classify the e-commerce transaction risks, and at the same time, according to the extracted joint feature variables, comprehensively evaluate the comprehensive risk level of the transaction.

[0011] The beneficial effects of the present invention: In the present invention, various key data of goods are obtained in real time from four parties: the e-commerce platform, merchants, payment institutions, and logistics, covering multi-dimensional information such as transactions, operations, payments, and logistics, and comprehensively covering all links of e-commerce transactions. Compared with a single data source, it can capture potential risks more comprehensively and timely, avoid risk misjudgment or missed judgment caused by information missing, and provide a solid data basis for risk prevention and control. In the present invention, joint feature variables such as the merchant transaction fluctuation index, the logistics timeliness deviation rate, the payment anomaly coefficient, and the business indicator deviation degree are extracted through scientific algorithms, and the risk status is deeply characterized from multiple angles such as merchant transaction dynamics, logistics timeliness changes, payment anomalies, and business indicator fluctuations. These feature variables can sensitively reflect the abnormal changes in the transaction process, making the risk assessment more targeted and accurate, and effectively identifying risk points that are difficult to discover by traditional methods. In the present invention, the intelligent early warning module and the risk assessment module respectively classify and comprehensively evaluate the e-commerce transaction risks from different dimensions. The intelligent early warning module makes a quick risk determination based on the threshold conditions of the joint feature variables, and the risk assessment module obtains the comprehensive risk value and classifies the level through the normalization processing, weight assignment, and comprehensive calculation of the joint feature variables. The dual assessment mechanisms complement and verify each other, greatly improving the accuracy and reliability of risk judgment and avoiding the limitations of a single assessment method. In the present invention, different levels of warning signals are triggered according to different risk levels, and risk information is accurately pushed to relevant responsible parties. The first-level risk signal is only sent to the e-commerce platform operators for easy timely attention; the second-level risk signal notifies both the operators and merchants simultaneously to promote collaborative handling between the two parties; the third-level risk signal sends an emergency warning to multiple parties including operators, payment institutions, and logistics companies to achieve cross-departmental joint prevention and control. This hierarchical warning mechanism makes risk response more efficient and orderly, and can take targeted measures in the shortest time to reduce risk losses. In the present invention, joint feature variable intervals are determined and weights are assigned based on historical data, and the evaluation model can be continuously optimized as data accumulates and business develops. Through the analysis and learning of a large amount of data, the risk assessment criteria and weight coefficients are continuously adjusted to make the risk control strategy more in line with the actual business needs, realizing the dynamic optimization and self-improvement of the risk control system, and enhancing the overall risk prevention and control ability of the e-commerce platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below with reference to the accompanying drawings.

[0013] Figure 1 It is a system block diagram of an intelligent risk control system for e-commerce digitization according to the present invention.

[0014] Figure 2 It is a schematic flowchart of an intelligent risk control method for e-commerce digitization according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Embodiment 1: Please refer to Figure 1 and Figure 2 As shown, the present invention is an intelligent risk control system for e-commerce digitization, including: A data acquisition module for real-time acquisition of various key data of goods from four parties: e-commerce platform, merchant, payment institution, and logistics; The various key data include: Transaction data of the e-commerce platform, specifically covering: order ID, user ID, order amount, order timestamp, user order IP address; Business operation data of the merchant, specifically covering: merchant ID, product category code, product price, inventory quantity, historical transaction volume and amount; The payment data of the payment institution specifically includes: payment order ID, payment channel code, payment amount, payment time, payment response status code; Among them, the payment response status code is a digital code returned by the payment gateway, bank or third-party payment platform after processing the transaction request, which is used to clearly indicate the processing status and result of the payment transaction; The processing status and result of the payment transaction specifically refer to the situation of identifying successful payment, failed payment or abnormal payment response; The logistics data of logistics specifically includes: waybill ID, pickup time, transportation node information, delivery time, logistics service provider code; The feature extraction module is used to extract joint feature variables for characterizing the risk situation based on various key data; The joint feature variables include merchant transaction volatility index, logistics timeliness deviation rate, payment anomaly coefficient, and business indicator deviation degree; Step A1, Merchant Transaction Volatility Index: Based on the time sequence relationship between the order timestamp in the transaction data and the current system time, two-dimensional order quantity statistics are carried out as follows: Count the order quantity within a preset standard observation window pushed forward from the current time node and mark it as D1; Count the average order quantity of each preset standard observation window within a preset specified period in the past and mark it as DP; In this embodiment, the standard observation window refers to 1 hour, and the specified period refers to 7 days; Through: ; Calculate the merchant transaction volatility index M; Among them, the merchant transaction volatility index measures the fluctuation range of the order quantity by comparing the difference between the order quantity of the merchant in a short period of time and the recent average level; If the fluctuation amplitude is too large, it may imply abnormal transaction behaviors such as brushing orders or malicious batch ordering; Step A2, Logistics Timeliness Deviation Rate: Select a waybill ID, and based on the time difference between the pickup time and the delivery time in the logistics data corresponding to the waybill ID, determine the actual delivery time of the waybill ID and mark it as P1; Based on the user's order placement IP address corresponding to the commodity of the waybill ID, obtain multiple historical waybill IDs in the logistics data corresponding to the same user's order placement IP address, and obtain the time difference between the pickup time and the delivery time among them, calculate the average delivery time of multiple historical waybills and mark it as PP; Through: ; Calculate the logistics timeliness deviation rate L; Among them, the logistics timeliness deviation rate compares the actual delivery duration with the industry average level to detect whether there are abnormal delays in the logistics link; Logistics delays may be caused by various risk factors, such as false logistics information, loss or detention of goods, etc.; Step A3, Payment anomaly coefficient: Based on the payment time in the payment data, obtain all payment order IDs within a specified observation period, and record the number of all payment order IDs as the total number of payments Z0; At the same time, based on the payment response status code, count the payment order IDs with unsuccessful payments, that is, the payment order IDs with payment failure or abnormal response situations; and record all payment order IDs with unsuccessful payments as the number of payment anomalies Z1. By: ; Calculate the payment anomaly coefficient Y; Among them, by statistically analyzing the proportion of failed and abnormal response situations in the total number of payments during the payment process, the stability of the payment link is evaluated; An overly high payment anomaly coefficient may indicate risks such as payment fraud and payment system vulnerabilities; Step A4, Deviation degree of business indicators: Based on the historical number of transactions in the business data, combined with the time series relationship of the current system time, count the number of transactions within a preset standard period pushed forward from the current time node, and mark it as B0; Count the number of transactions in each preset standard period within the past preset observation period, and mark it as Bt, where t = 1, 2,..., k, and k represents the number of standard periods within the specified period; In this embodiment, the standard period is defined as 1 day, and the observation period is defined as 30 days; By: ; Calculate the deviation degree of business indicators H; Among them, the historical number of transactions is a business indicator. Through this deviation degree of business indicators, it can be judged whether the business behavior of the merchant has abnormal fluctuations in the short term. The greater the deviation degree, the higher the risk possibility; The intelligent early warning module is used to classify the e-commerce transaction risks according to the combined feature variables; When M ∈ [10, 20), and at the same time L < 10%, Y < 5% and H < 1.5 all hold, then it is determined that the e-commerce transaction risk is a first-level risk; The first-level risk indicates that there are slight abnormal fluctuations in the transaction, but the overall risk level is relatively low, and it is necessary to continuously monitor the changes in subsequent transaction data; When any of the conditions M ∈ [20, 50), L ∈ [10%, 30%), Y ∈ [5%, 15%), H ∈ [1.5, 3) is satisfied, the e-commerce transaction risk is determined to be a secondary risk; The secondary risk indicates that there are relatively obvious abnormal signs in the transaction, which may involve potential risk factors and certain measures need to be taken for investigation and monitoring; When at least one of the conditions M ≥ 50, L ≥ 30%, Y ≥ 15%, H ≥ 3 holds, the e-commerce transaction risk is determined to be a tertiary risk; The tertiary risk indicates that there is a high risk in the transaction, which may involve serious problems such as fraud and malicious brushing of orders, and an emergency disposal process needs to be started immediately.

[0017] Embodiment 1 constructs the basic framework of the e-commerce digital intelligent risk control system. The key data covering multiple dimensions such as transactions, operations, payments, and logistics are obtained in real time from four parties: e-commerce platforms, merchants, payment institutions, and logistics through the data acquisition module, providing comprehensive and accurate data support for risk control analysis. The feature extraction module extracts joint feature variables such as the merchant transaction fluctuation index, logistics timeliness deviation rate, payment anomaly coefficient, and business indicator deviation degree based on these data, depicting the risk situation from different angles. The intelligent early warning module classifies the e-commerce transaction risk according to the joint feature variables, can quickly identify the risk level in the transaction, and issue an early warning in a timely manner when abnormal fluctuations occur. This embodiment realizes the systematic and quantitative assessment of e-commerce transaction risks, helps e-commerce platforms discover potential risks in a timely manner, provides a powerful technical means for subsequent risk prevention and management, and effectively guarantees the security and stability of e-commerce transactions.

[0018] Embodiment 2: Please refer to Figure 1 and Figure 2 As shown in the figure, as Embodiment 2 of the present invention, in the specific implementation of this application, compared with Embodiment 1, the technical solution of this embodiment is only different from that of Embodiment 1 in that this embodiment further includes: A risk assessment module for comprehensively evaluating the comprehensive risk level of the transaction according to the extracted joint feature variables; The specific method is as follows: Step B1. Normalization processing of joint feature variables: In historical data, multiple groups of various key data similar to the commodity are extracted. Then, the corresponding joint feature variables are extracted from the multiple groups of various key data in the historical data through the feature extraction module. Then, the minimum value and the maximum value of each joint feature variable are extracted therefrom, and a joint feature variable interval is formed; The joint feature variable interval covers the merchant transaction fluctuation interval [M min , M max , the logistics timeliness deviation rate interval [L min , L max, Payment Abnormality Interval [Y min , Y max and Business Indicator Deviation Interval [H min , H max ; For the merchant transaction volatility index M of the current commodity; By: ; Calculate the normalized merchant transaction volatility index M1; For the logistics timeliness deviation rate L of the current commodity; By: ; Calculate the normalized logistics timeliness deviation rate L1; For the payment abnormality coefficient Y of the current commodity; By: ; Calculate the normalized payment abnormality coefficient Y1; For the business indicator deviation H of the current commodity; By: ; Calculate the normalized business indicator deviation H1; Step B2, Weight Extraction of Each Joint Feature Variable: Extract the corresponding weight coefficients assigned in advance for each joint feature variable; Specifically as follows: w1: Weight coefficient of the merchant transaction volatility index; w2: Weight coefficient of the logistics timeliness deviation rate; w3: Weight coefficient of the payment abnormality coefficient; w4: Weight coefficient of the business indicator deviation; Among them, the determination of the weight coefficients is based on the actual experience of the e-commerce industry and the analysis of the importance of various risk factors; the value range of the weights is 0 - 1, and w1 + w2 + w3 + w4 = 1; For example, through the analysis of a large amount of historical transaction data and the evaluation of industry experts, it is found that problems in the payment link may lead to direct financial losses and are relatively more critical. Therefore, the weight w3 assigned to the payment abnormality coefficient Y may be relatively high; while the merchant transaction volatility index M is also important, but its urgency is slightly lower than that of the payment abnormality situation, and the weight w1 assigned to it is relatively lower; Step B3, Comprehensive Risk Value Calculation: By: ; Calculate the comprehensive risk value R; Step B4, Comprehensive Risk Level Assessment: According to the comprehensive risk value R, the risk levels are divided into three levels: low risk, medium risk, and high risk; The specific classification criteria are as follows: When 0 ≤ R < 0.3, the comprehensive risk level is determined to be the low-risk level; The low-risk level indicates that the possibility of abnormalities in each link of the transaction process is relatively low, and the overall transaction is relatively safe and reliable; When 0.3 ≤ R < 0.6, the comprehensive risk level is determined to be the medium-risk level; The medium-risk level indicates that there are certain risks in the transaction, and something abnormal may have occurred in one or some links, which requires further attention and review; When 0.6 ≤ R ≤ 1, the comprehensive risk level is determined to be the high-risk level; The high-risk level indicates that there are relatively large risks in the transaction, and there are likely to be bad behaviors such as fraud and illegal operations, and immediate measures need to be taken for handling, such as suspending the transaction and further investigation.

[0019] Based on Embodiment 1, Embodiment 2 adds a risk assessment module. By jointly performing normalization processing of feature variables, weight extraction, and calculation of the comprehensive risk value, the comprehensive risk level of the transaction is evaluated. This comprehensive evaluation method takes into account the differences in the importance of different risk factors in e-commerce transactions. By assigning weight coefficients, the risk assessment becomes more scientific and reasonable. Compared with the single risk level classification, Embodiment 2 can more comprehensively and accurately reflect the risk situation of the transaction, provide a more valuable risk assessment result for the e-commerce platform, help the platform make more accurate decisions in risk control, take more effective risk response measures, and further improve the ability and level of e-commerce transaction risk prevention and control.

[0020] Embodiment 3: Please refer to Figure 1 and Figure 2 As shown, as Embodiment 3 of the present invention, in the specific implementation of this application, compared with Embodiment 1 and Embodiment 2, the technical solution of this embodiment is to combine and implement the solutions of the above Embodiment 1 and Embodiment 2. The difference between the technical solution of this embodiment and Embodiment 1 and Embodiment 2 is only that in this embodiment, different levels of warning signals are triggered according to the results obtained by the intelligent warning module and the risk assessment module; When the e-commerce transaction risk is determined to be a first-level risk or the comprehensive risk level is determined to be the low-risk level, a first-level risk signal is generated, and the first-level risk signal is used to send a risk reminder message to the e-commerce platform operators; It is informed in the form of an email or a system message to prompt daily attention and tracking of the transaction data of relevant merchants. There is no need to take immediate intervention actions, but the data change situation needs to be reported regularly; When the e-commerce transaction risk is determined to be a secondary risk or the comprehensive risk level is determined to be a medium risk level, a secondary risk signal is generated, and the secondary risk signal is used to send a warning notice to the e-commerce platform operators and merchants; The e-commerce platform operators receive SMS and email notifications, and are required to conduct a preliminary investigation on the merchant, such as checking the recent business strategy adjustments of the merchant, whether there are new products launched, etc.; the merchant receives an in-site message notification, prompting it to conduct a self-inspection of the transaction orders, check for any abnormal operations, and feedback the self-inspection results to the platform within the specified time; When the e-commerce transaction risk is determined to be a tertiary risk or the comprehensive risk level is determined to be a high risk level, a tertiary risk signal is generated, and the tertiary risk signal is used to send an emergency warning to the e-commerce platform operators, the risk control department of the payment institution, and the risk monitoring department of the logistics company; The e-commerce platform operators need to freeze the merchant's account within 1 hour and suspend all its trading activities; after receiving the notice, the risk control department of the payment institution starts the procedures for freezing and investigating the funds involved in the risky transactions and delays the settlement of related transactions; the risk monitoring department of the logistics company conducts key monitoring on the goods of the relevant waybills, and can suspend the transportation and conduct a goods inspection if necessary to prevent the goods from being illegally received.

[0021] Example 3 combines the risk level classification in Example 1 with the comprehensive risk assessment in Example 2, and triggers different levels of warning signals according to different risk determination results. This hierarchical warning mechanism formulates differentiated response strategies for different risk levels, from the daily attention to primary risks, to the preliminary investigation and self-inspection of secondary risks, and then to the emergency handling of tertiary risks, realizing the refinement and dynamicization of risk control. By sending warning signals of corresponding levels to relevant parties such as e-commerce platform operators, merchants, payment institutions, and logistics companies, it clarifies the responsibilities and action plans of all parties in risk prevention and control, improves the efficiency and coordination of risk handling, effectively reduces the losses that may be brought by e-commerce transaction risks, and ensures the healthy operation of the e-commerce transaction ecosystem.

[0022] Example 4: Please refer to Figure 1 and Figure 2 As shown in, as Example 4 of the present invention, in the specific implementation of this application, compared with Example 1, Example 2, and Example 3, the technical solution of this example lies in the combined implementation of the solutions of the above Example 1, Example 2, Example 3, and Example 4.

[0023] Embodiment 4 integrates the technical solutions of the first three embodiments to form a complete, comprehensive, and in-depth e-commerce digital intelligent risk control system. This system can not only extract risk features from multi-source data, conduct grading and comprehensive evaluation, but also implement hierarchical early warning and disposal according to different risk results. Through multi-dimensional data collection, scientific risk analysis models, and perfect early warning response mechanisms, Embodiment 4 realizes the full-process and all-round control of e-commerce transaction risks, minimizes transaction risks, safeguards the legitimate rights and interests of e-commerce platforms, merchants, and consumers, improves the overall risk resistance of the e-commerce industry, and lays a solid technical foundation for the stable development of e-commerce business and the healthy prosperity of the industry.

[0024] The present invention also provides an intelligent risk control method for e-commerce digitization, which is implemented through an intelligent risk control system for e-commerce digitization. The method includes the following steps: First step, obtaining data: obtaining various key data of goods in real time; Second step, refining features: refining joint feature variables based on various key data; Third step, grade evaluation: grading the e-commerce transaction risks based on the joint feature variables, and comprehensively evaluating the comprehensive risk grade of the transaction according to the refined joint feature variables.

[0025] It should be stated that: all data collected in this application are collected with the consent and authorization of users, and the uses of all data are legal and compliant, and the use and processing of data comply with the relevant laws, regulations, and standards of the relevant regions.

[0026] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the actual situation. The preset parameters and threshold selections in the formulas are set by those skilled in the art according to the actual situation.

[0027] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

Claims

1. An intelligent risk control system for e-commerce digitization, characterized in that, It includes: A data acquisition module for real-time acquisition of various key data of goods; A feature extraction module for extracting combined feature variables for characterizing the risk situation based on various key data. The combined feature variables include the merchant transaction fluctuation index, the logistics timeliness deviation rate, the payment anomaly coefficient, and the business indicator deviation degree; An intelligent early warning module for classifying the levels of e-commerce transaction risks based on the combined feature variables; A risk assessment module for comprehensively evaluating the comprehensive risk level of a transaction according to the extracted combined feature variables.

2. The intelligent risk control system for e-commerce digitization according to claim 1, characterized in that, The various key data include transaction data obtained from an e-commerce platform, which covers the order ID, the order placement timestamp, and the user's order placement IP address; business operation data obtained from a merchant platform, which covers the historical transaction volume; payment data obtained from a payment institution, which covers the payment order ID, the payment time, and the payment response status code; and logistics data from a logistics platform, which covers the waybill ID, the pickup time, and the delivery time; Among them, the payment response status code is used to clearly indicate the processing status and result of a payment transaction; the processing status and result of a payment transaction specifically refer to the situation of identifying successful payment, failed payment, or payment anomaly response.

3. An intelligent risk control system for e-commerce digitization according to claim 2, characterized in that, The extraction method of the merchant transaction fluctuation index is as follows: Based on the chronological relationship between the order timestamp in the transaction data and the current system time, perform two-dimensional order quantity statistics, specifically as follows: Statistical the order quantity within a preset standard observation window pushed forward from the current time node, and mark it as D1; Statistical the average order quantity of each preset standard observation window within a preset specified period in the past, and mark it as DP; Passed by: , the merchant transaction volatility index M is calculated.

4. An intelligent risk control system for e-commerce digitization according to claim 3, characterized in that, The extraction method of the logistics timeliness deviation rate is as follows: Select a waybill ID, and based on the time difference between the pickup time and the delivery time in the logistics data corresponding to the waybill ID, determine the actual delivery time of the waybill ID, and mark it as P1; Based on the user's order placement IP address corresponding to the goods of the waybill ID, obtain multiple historical waybill IDs in the logistics data corresponding to the same user's order placement IP address, and obtain the time difference between the pickup time and the delivery time from them, calculate the average delivery time of multiple historical waybills, and mark it as PP; Passed by: , the logistics timeliness deviation rate L is calculated.

5. An intelligent risk control system for e-commerce digitization according to claim 4, characterized in that, The extraction method of the payment anomaly coefficient is as follows: Based on the payment time in the payment data, obtain all payment order IDs within a specified observation period, and record the number of all payment order IDs as the total number of payments Z0; At the same time, based on the payment response status code, count the payment order IDs with unsuccessful payments, that is, the payment order IDs with payment failure or payment anomaly response situations; and record all payment order IDs with unsuccessful payments as the number of payment anomalies Z1; Passed by: , the payment anomaly coefficient Y is calculated.

6. An intelligent risk control system for e-commerce digitization according to claim 5, characterized in that, The extraction method of the business indicator deviation degree is as follows: Based on the historical transaction volume in the business operation data and combined with the chronological relationship of the current system time, count the number of transactions within a preset standard time period pushed forward from the current time node, and mark it as B0; Statistical the number of transactions in each preset standard time period within a preset observation period in the past; Adopted by: , the deviation degree H of the operation index is calculated; In the formula, BP and BB are the average value and standard deviation of the corresponding transaction volumes in each preset standard time period within a preset observation period in the past, respectively.

7. An intelligent risk control system for e-commerce digitization according to claim 6, characterized in that, The level classification method of the intelligent early warning module is as follows: When M∈[Ma,Mb), and L<La, Y<Ya and H<Ha all hold, the e-commerce transaction risk is determined to be a first-level risk; When any of the following conditions is met: M∈[Mb,Mc), L∈[La,Lb), Y∈[Ya,Yb), H∈[Ha,Hb), the e-commerce transaction risk is determined to be a secondary risk; When at least one of the following conditions is true: M≥Mc, L≥Lb, Y≥Yb, H≥Hb, the e-commerce transaction risk is determined to be level 3 risk; Among them: Ma, Mb, Mc are thresholds pre-set based on the merchant transaction fluctuation index; La, Lb are thresholds pre-set based on the logistics timeliness deviation rate; Ya, Yb are thresholds pre-set based on the payment anomaly coefficient; Ha, Hb are thresholds pre-set based on the deviation of operating indicators.

8. An intelligent risk control system for e-commerce digitization according to claim 6, characterized in that, The comprehensive risk level is assessed as follows: In the historical data, multiple groups of key data similar to the product are extracted, and then the corresponding joint feature variables are extracted from the multiple groups of key data in the historical data through the feature extraction module, and then the minimum and maximum values ​​of each joint feature variable are extracted from them to form a joint feature variable interval; Then, combined with the interval of the joint characteristic variables, and through the normalization processing formula, the normalized merchant transaction fluctuation index M1, logistics timeliness deviation rate L1, payment anomaly coefficient Y1, and business indicator deviation H1 are calculated; Extract the corresponding weight coefficients pre-assigned according to the merchant transaction fluctuation index, logistics timeliness deviation rate, payment anomaly coefficient, and business indicator deviation, and mark them as w1, w2, w3, and w4 in sequence; among which w1+w2+w3+w4=1; Adopted by: , the comprehensive risk value R is calculated; According to the comprehensive risk value R, the risk level is divided into three levels: low risk, medium risk and high risk.

9. An intelligent risk control system for e-commerce digitization according to claim 8, characterized in that, The classification criteria based on the comprehensive risk value are as follows: When Ra≤R<Rb, the comprehensive risk level is judged to be a low risk level; When Rb≤R<Rc, the comprehensive risk level is determined to be medium risk level; When Rc≤R≤Rd, the comprehensive risk level is determined to be a high risk level; Among them, Ra, Rb, Rc, and Rd are preset comprehensive risk thresholds, and Ra<Rb<Rc<Rd.

10. An intelligent risk control method for e-commerce digitization, which is implemented by an intelligent risk control system for e-commerce digitization according to any one of claims 1-9, characterized in that, The method comprises the following steps: Get data: Get all kinds of key data of goods in real time; Extract features: Extract joint feature variables based on various key data; Level assessment: E-commerce transaction risks are classified into different levels based on joint characteristic variables, and the comprehensive risk level of the transaction is comprehensively assessed based on the extracted joint characteristic variables.

Images