Security protection method and system for e-commerce platform

Abstract

The invention discloses a security protection method and system for an e-commerce platform, belongs to the technical field of e-commerce, and aims to solve the problem that online payment fraud is increasingly severe, and the method comprises the steps of collecting and analyzing historical transaction data, constructing a risk assessment model based on the historical transaction data, and establishing a risk assessment model based on the risk assessment model. Transaction behaviors on the e-commerce platform are monitored in real time, whether the transaction behaviors are abnormal behaviors is judged, a real-time transaction verification mechanism is introduced based on the abnormal behaviors, and the confirmed fraudulent behaviors are analyzed and traced; through the risk assessment model and real-time transaction verification, fraudulent transactions are effectively identified and prevented, user fund safety is protected, transaction safety is improved, the intelligent monitoring and early warning system can timely find and early warn abnormal transaction behaviors, fraudulent behaviors are reduced, fraudulent risks are reduced, and the safety of the transaction is improved while the safety is ensured. Interference on normal transactions is reduced as much as possible, and user experience is improved and optimized.

Description

Technical Field

[0001] The present invention belongs to the technical field of e-commerce, and in particular relates to a security protection method and system for an e-commerce platform. Background Art

[0002] An e-commerce platform is a platform for enterprises or individuals to conduct online transactions and negotiations. The enterprise e-commerce platform is built on the Internet. It is not only a virtual network space for business activities, but also a management environment to ensure the smooth operation of business. This platform coordinates and integrates information flow, goods flow, and capital flow, so that they can flow in an orderly, relevant, and efficient manner. Enterprises and merchants can make full use of the shared resources such as network infrastructure, payment platform, security platform, and management platform provided by the e-commerce platform to carry out their own business activities effectively and at low cost. The construction of the e-commerce platform aims to promote the development of e-commerce. It not only realizes the function of primary online shopping, but also builds a secure and easily expandable business framework system, supporting B2B (business to business), B2C (business to consumer), C2C (consumer to consumer), O2O (online to offline) and other application environments. These platforms display, promote or sell goods through the Internet, providing small and medium-sized enterprises and individuals with opportunities for self-employment and independent marketing, and achieving the goal of rapid profitability.

[0003] With the rapid development of e-commerce, online payment has become one of the main ways for consumers to shop. The incidence of e-commerce fraud has shown an upward trend in recent years. The resulting online payment fraud problem is also becoming increasingly serious, including false transactions, stolen credit cards, phishing websites and other fraud methods, which seriously threaten the financial security of users and the reputation of e-commerce platforms.

[0004] Therefore, there is a need for a security protection method and system for an e-commerce platform to solve the increasingly serious problem of online payment fraud in the prior art. Summary of the invention

[0005] The purpose of the present invention is to provide a security protection method and system for an e-commerce platform to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a security protection method for an e-commerce platform, comprising:

[0007] Collect and analyze historical transaction data:

[0008] Build risk assessment models based on historical transaction data;

[0009] Monitor transaction behaviors on e-commerce platforms in real time and determine whether they are abnormal behaviors;

[0010] Introducing real-time transaction verification mechanisms based on abnormal behavior;

[0011] Analyze and trace confirmed fraudulent activities.

[0012] It should be noted in the plan that the collection and analysis of historical transaction data includes:

[0013] Collect historical transaction data on e-commerce platforms. This data should include various transaction characteristics, including but not limited to transaction amount, transaction time, transaction location, payment method, user historical behavior records, and device information;

[0014] Preprocess the data, including converting transaction records into a form suitable for CNN processing, encoding transaction features into a fixed-length matrix, and normalizing numerical data to ensure that all features are on the same scale;

[0015] The preprocessed dataset is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters and perform feature selection, and the test set is used to evaluate the final performance of the model.

[0016] It is further worth noting that the construction of the risk assessment model based on historical transaction data includes designing a CNN model framework and a training model, wherein the design of the CNN model framework includes:

[0017] Input layer: Define the input layer according to the shape of the preprocessed data. The data is encoded as a two-dimensional matrix, so the input layer should match the size of the matrix.

[0018] Convolutional Layer:

[0019]

[0020] Among them, k is the size of the convolution kernel, Input is the input data, Kernel is the convolution kernel, and Output is the output feature map;

[0021] Activation function: Apply nonlinear activation function after convolutional layer to increase the nonlinear ability of the network

[0022] ReLU(x)=max(0,x)

[0023] Pooling layer: Use the pooling layer to reduce the dimension of the feature map, reduce the amount of calculation, and extract the most important features;

[0024] Fully connected layer: After the convolutional layer and the pooling layer, a fully connected layer is added to integrate all local features and output the final prediction result, which is set as the fraud probability;

[0025] Output layer: Use the sigmoid function to generate the final prediction probability

[0026]

[0027] It should be further explained that the training model includes:

[0028] The cross entropy loss function is used to measure the difference between the model prediction and the true label:

[0029]

[0030] Where N is the number of samples, y i is the true label, is the predicted probability.

[0031] The loss is calculated through forward propagation, and then the weights are updated through backpropagation, and this process is repeated until the stopping condition is met.

[0032] As a preferred implementation, the real-time monitoring of transaction behaviors on the e-commerce platform and determining whether the behaviors are abnormal include:

[0033] Deploy the trained risk assessment model into the actual production environment to analyze new transaction records in real time to detect potential fraud.

[0034] The e-commerce platform sets a monitoring threshold based on the user's historical usage habits. When the predicted probability is higher than the threshold, the transaction behavior is judged to be abnormal. The system automatically triggers the early warning mechanism and notifies the user and platform administrator via email, SMS, APP push, etc.

[0035] The early warning mechanism supports custom rule settings, allowing the platform to adjust monitoring thresholds and early warning strategies according to actual needs.

[0036] As a preferred implementation, the introduction of a real-time transaction verification mechanism based on abnormal behavior includes:

[0037] Introducing biometric technology and secondary verification methods during the transaction process to ensure the authenticity and legality of the transaction:

[0038] For transactions identified as abnormal behavior, the system automatically requires users to go through additional identity verification steps to reduce the risk of fraud;

[0039] The real-time verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security.

[0040] As a preferred implementation, the analysis and tracing of confirmed fraudulent behavior includes:

[0041] Conduct in-depth analysis of confirmed fraudulent behaviors, extract fraud characteristics, and optimize risk assessment models;

[0042] Establish a fraud behavior database to record the behavior patterns and characteristics of fraudsters and provide data support for future fraud detection;

[0043] Provide fraud tracing function to help the platform quickly locate the source of fraud and take corresponding measures.

[0044] A security protection system for an e-commerce platform is further proposed, which is used to implement the security protection method for an e-commerce platform as described above, including:

[0045] The data collection module is used to collect historical transaction data on the e-commerce platform. These data should include various characteristics of the transactions;

[0046] The data processing module is used to pre-process the data, including converting transaction records into a form suitable for CNN processing, encoding transaction features into a matrix of fixed length; normalizing numerical data to ensure that all features are on the same scale; and dividing the pre-processed data set into training set, validation set, and test set;

[0047] Risk assessment module, used to build risk assessment models, model evaluation and adjustment;

[0048] The monitoring and early warning module includes a monitoring unit, a judgment unit and an early warning unit. The monitoring unit is used to monitor the transaction behaviors on the e-commerce platform in real time, input new transaction behaviors into the risk assessment module, and obtain the predicted probability; the judgment unit judges whether the obtained predicted probability is higher than the threshold. Transaction behaviors higher than the threshold are abnormal behaviors, and transaction behaviors lower than the threshold are normal behaviors: the early warning unit, abnormal behaviors are used by the system to automatically trigger the early warning mechanism, and notify users and platform administrators through emails, SMS, APP push, etc.;

[0049] The transaction verification module is used to introduce biometric technology and secondary verification methods in the transaction process to ensure the authenticity and legality of the transaction. The verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security;

[0050] The fraud analysis module includes an analysis unit and a tracing unit. The analysis unit conducts in-depth analysis of confirmed fraudulent behaviors, extracts fraud characteristics, optimizes risk assessment models, establishes a fraudulent behavior database, and records the behavior patterns and characteristics of fraudsters. The tracing unit provides a fraudulent behavior tracing function to help the platform quickly locate the source of fraud and take corresponding measures.

[0051] Compared with the prior art, the security protection method and system for an e-commerce platform provided by the present invention have at least the following beneficial effects:

[0052] Through the established risk assessment model and real-time transaction verification, fraudulent transactions can be effectively identified and prevented, user funds can be protected, and transaction security can be improved; the intelligent monitoring and early warning system can timely detect and warn of abnormal transaction behaviors, reduce the occurrence of fraudulent behaviors, and reduce fraud risks; while ensuring security, minimize interference with normal transactions, improve user experience, and optimize user experience; through effective security protection measures, enhance the credibility and competitiveness of the e-commerce platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the process flow of the security protection method for an e-commerce platform of the present invention;

[0054] Figure 2 The figure is a structural block diagram of the security protection system for the e-commerce platform of the present invention. DETAILED DESCRIPTION

[0055] The present invention will be further described below in conjunction with the embodiments.

[0056] Reference Figure 1 As shown, the present invention provides a security protection method for an e-commerce platform, comprising:

[0057] Collect and analyze historical transaction data:

[0058] Build risk assessment models based on historical transaction data;

[0059] Monitor transaction behaviors on e-commerce platforms in real time and determine whether they are abnormal behaviors;

[0060] Introducing real-time transaction verification mechanisms based on abnormal behavior;

[0061] Analyze and trace confirmed fraudulent activities.

[0062] Wherein, the collection and analysis of historical transaction data includes:

[0063] Collect historical transaction data on e-commerce platforms. This data should include various transaction characteristics, including but not limited to transaction amount, transaction time, transaction location, payment method, user historical behavior records, and device information;

[0064] Preprocess the data, including converting transaction records into a form suitable for CNN processing, encoding transaction features into a fixed-length matrix, and normalizing numerical data to ensure that all features are on the same scale;

[0065] The preprocessed dataset is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters and perform feature selection, and the test set is used to evaluate the final performance of the model.

[0066] It can be understood that the construction of the risk assessment model based on historical transaction data includes designing a CNN model framework and a training model, wherein the design of the CNN model framework includes:

[0067] Input layer: Define the input layer according to the shape of the preprocessed data. The data is encoded as a two-dimensional matrix, so the input layer should match the size of the matrix.

[0068] Convolutional Layer:

[0069]

[0070] Among them, k is the size of the convolution kernel, Input is the input data, Kernel is the convolution kernel, and Output is the output feature map;

[0071] Activation function: Apply nonlinear activation function after convolutional layer to increase the nonlinear ability of the network

[0072] ReLU(x)=max(0,x)

[0073] Pooling layer: Use the pooling layer to reduce the dimension of the feature map, reduce the amount of calculation, and extract the most important features;

[0074] Fully connected layer: After the convolutional layer and the pooling layer, a fully connected layer is added to integrate all local features and output the final prediction result, which is set as the fraud probability;

[0075] Output layer: Use the sigmoid function to generate the final prediction probability

[0076]

[0077] Wherein, the training model includes:

[0078] The cross entropy loss function is used to measure the difference between the model prediction and the true label:

[0079]

[0080] Where N is the number of samples, y i is the true label, is the predicted probability.

[0081] The loss is calculated through forward propagation, and then the weights are updated through backpropagation, and this process is repeated until the stopping condition is met.

[0082] Among them, convolutional neural networks (CNNs) are a special type of deep feedforward neural networks that have a wide range of applications in image processing, video analysis, natural language processing (especially in tasks such as text classification), and speech recognition. They are special because they are able to exploit the spatial hierarchy of the input data, which makes them very effective when processing data with obvious spatial structure, such as images and videos.

[0083] It is understandable that the real-time monitoring of transaction behaviors on the e-commerce platform and determining whether it is abnormal behavior includes:

[0084] Deploy the trained risk assessment model into the actual production environment to analyze new transaction records in real time to detect potential fraud.

[0085] The e-commerce platform sets a monitoring threshold based on the user's historical usage habits. When the predicted probability is higher than the threshold, the transaction behavior is judged to be abnormal. The system automatically triggers the early warning mechanism and notifies the user and platform administrator via email, SMS, APP push, etc.

[0086] The early warning mechanism supports custom rule settings, allowing the platform to adjust monitoring thresholds and early warning strategies according to actual needs.

[0087] It is understandable that the real-time transaction verification mechanism introduced based on abnormal behavior includes:

[0088] Introducing biometric technology and secondary verification methods during the transaction process to ensure the authenticity and legality of the transaction:

[0089] For transactions identified as abnormal behavior, the system automatically requires users to go through additional identity verification steps to reduce the risk of fraud;

[0090] The real-time verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security.

[0091] The analysis and tracing of confirmed fraudulent activities include:

[0092] Conduct in-depth analysis of confirmed fraudulent behaviors, extract fraud characteristics, and optimize risk assessment models;

[0093] Establish a fraud behavior database to record the behavior patterns and characteristics of fraudsters and provide data support for future fraud detection;

[0094] Provide fraud tracing function to help the platform quickly locate the source of fraud and take corresponding measures.

[0095] For further information, please refer to Figure 2As shown in the figure, a security protection method for an e-commerce platform is proposed. This scheme proposes a security protection system for an e-commerce platform, including:

[0096] The data collection module is used to collect historical transaction data on the e-commerce platform. These data should include various characteristics of the transactions;

[0097] The data processing module is used to pre-process the data, including converting transaction records into a form suitable for CNN processing, encoding transaction features into a matrix of fixed length; normalizing numerical data to ensure that all features are on the same scale; and dividing the pre-processed data set into training set, validation set, and test set;

[0098] Risk assessment module, used to build risk assessment models, model evaluation and adjustment;

[0099] The monitoring and early warning module includes a monitoring unit, a judgment unit and an early warning unit. The monitoring unit is used to monitor the transaction behaviors on the e-commerce platform in real time, input new transaction behaviors into the risk assessment module, and obtain the predicted probability; the judgment unit judges whether the obtained predicted probability is higher than the threshold. Transaction behaviors higher than the threshold are abnormal behaviors, and transaction behaviors lower than the threshold are normal behaviors: the early warning unit, abnormal behaviors are used by the system to automatically trigger the early warning mechanism, and notify users and platform administrators through emails, SMS, APP push, etc.;

[0100] The transaction verification module is used to introduce biometric technology and secondary verification methods in the transaction process to ensure the authenticity and legality of the transaction. The verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security;

[0101] The fraud analysis module includes an analysis unit and a tracing unit. The analysis unit conducts in-depth analysis of confirmed fraudulent behaviors, extracts fraud characteristics, optimizes risk assessment models, establishes a fraudulent behavior database, and records the behavior patterns and characteristics of fraudsters. The tracing unit provides a fraudulent behavior tracing function to help the platform quickly locate the source of fraud and take corresponding measures.

[0102] The use process of the above-mentioned security protection system for an e-commerce platform is as follows:

[0103] Deploy the security protection system to the server of the e-commerce platform to ensure that the system can operate normally and integrate seamlessly with the e-commerce platform;

[0104] Configure various system parameters according to the specific needs and security strategies of the e-commerce platform, including the threshold of the risk assessment model and the trigger conditions of monitoring and early warning;

[0105] The monitoring unit monitors the transaction behaviors on the e-commerce platform in real time, inputs new transaction behaviors into the risk assessment module, and obtains the predicted probability;

[0106] The judgment unit judges whether the predicted probability is higher than the threshold. The transaction behavior higher than the threshold is abnormal behavior, and the transaction behavior lower than the threshold is normal behavior:

[0107] When abnormal behavior is detected, the early warning unit automatically triggers the early warning mechanism and notifies users and platform administrators via email, SMS, APP push, etc.

[0108] At the same time, the transaction verification module introduces biometric technology and secondary verification methods to improve transaction security;

[0109] The analysis unit conducts in-depth analysis of confirmed fraudulent behaviors, extracts fraud characteristics, optimizes risk assessment models, establishes a fraudulent behavior database, and records the behavior patterns and characteristics of fraudsters;

[0110] The tracing unit provides fraud tracing functions to help the platform quickly locate the source of fraud and take corresponding measures.

[0111] To sum up, the advantages of the present invention are: through the set risk assessment model and real-time transaction verification, fraudulent transactions can be effectively identified and prevented, user funds can be protected, and transaction security can be improved; the intelligent monitoring and early warning system can timely detect and warn of abnormal transaction behaviors, reduce the occurrence of fraudulent behaviors, and reduce the risk of fraud; while ensuring security, it minimizes interference with normal transactions, improves user experience, and optimizes user experience; through effective security protection measures, the credibility and competitiveness of the e-commerce platform are enhanced.

[0112] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A security protection method for an e-commerce platform, characterized in that: include: Collect and analyze historical transaction data: Build risk assessment models based on historical transaction data; Monitor transaction behaviors on e-commerce platforms in real time and determine whether they are abnormal behaviors; Introducing real-time transaction verification mechanisms based on abnormal behavior; Analyze and trace confirmed fraudulent activities.

2. A security protection method for an e-commerce platform according to claim 1, characterized in that: The collection and analysis of historical transaction data includes: Collect historical transaction data on e-commerce platforms. This data should include various transaction characteristics, including but not limited to transaction amount, transaction time, transaction location, payment method, user historical behavior records, and device information; Preprocess the data, including converting transaction records into a form suitable for CNN processing, encoding transaction features into a fixed-length matrix, and normalizing numerical data to ensure that all features are on the same scale; The preprocessed dataset is divided into training set, validation set and test set. The training set is used to train the model, the validation set is used to adjust the model parameters and perform feature selection, and the test set is used to evaluate the final performance of the model.

3. A security protection method for an e-commerce platform according to claim 2, characterized in that: The risk assessment model based on historical transaction data is constructed by designing a CNN model framework and a training model, wherein the design of the CNN model framework includes: Input layer: Define the input layer according to the shape of the preprocessed data. The data is encoded as a two-dimensional matrix, so the input layer should match the size of the matrix. Convolutional Layer: Among them, k is the size of the convolution kernel, Input is the input data, Kernel is the convolution kernel, and Output is the output feature map; Activation function: Apply a nonlinear activation function after the convolutional layer to increase the nonlinear ability of the network. ReLU(x) = max(0,x) Pooling layer: Use the pooling layer to reduce the dimension of the feature map, reduce the amount of calculation, and extract the most important features; Fully connected layer: After the convolutional layer and the pooling layer, a fully connected layer is added to integrate all local features and output the final prediction result, which is set as the fraud probability; Output layer: Use the sigmoid function to generate the final prediction probability 4. A security protection method for an e-commerce platform according to claim 3, characterized in that: The training model includes: The cross entropy loss function is used to measure the difference between the model prediction and the true label: Where N is the number of samples, y i is the true label, is the predicted probability. The loss is calculated through forward propagation, and then the weights are updated through backpropagation, and this process is repeated until the stopping condition is met.

5. A security protection method for an e-commerce platform according to claim 4, characterized in that: The real-time monitoring of transaction behaviors on the e-commerce platform and determining whether the behaviors are abnormal include: Deploy the trained risk assessment model to the actual production environment and perform real-time analysis on new transaction records to detect potential fraud. The e-commerce platform sets a monitoring threshold based on the user's historical usage habits. When the predicted probability is higher than the threshold, the transaction behavior is judged to be abnormal. The system automatically triggers the early warning mechanism and notifies the user and platform administrator via email, SMS, APP push, etc. The early warning mechanism supports custom rule settings, allowing the platform to adjust monitoring thresholds and early warning strategies according to actual needs.

6. A security protection method for an e-commerce platform according to claim 5, characterized in that: The real-time transaction verification mechanism introduced based on abnormal behavior includes: Introducing biometric technology and secondary verification methods during the transaction process to ensure the authenticity and legality of the transaction: For transactions identified as abnormal behavior, the system automatically requires users to take additional identity verification steps to reduce the risk of fraud; The real-time verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security.

7. A security protection method for an e-commerce platform according to claim 6, characterized in that: The analysis and tracing of confirmed fraudulent activities include: Conduct in-depth analysis of confirmed fraudulent behaviors, extract fraud characteristics, and optimize risk assessment models; Establish a fraud behavior database to record the behavior patterns and characteristics of fraudsters and provide data support for future fraud detection; Provide fraud tracing function to help the platform quickly locate the source of fraud and take corresponding measures.

8. A security protection system for an e-commerce platform is used to implement a security protection method for an e-commerce platform as claimed in any one of claims 1 to 7, characterized in that: include: The data collection module is used to collect historical transaction data on the e-commerce platform. These data should include various characteristics of the transactions; The data processing module is used to pre-process the data, including converting transaction records into a form suitable for CNN processing and encoding transaction features into a matrix of fixed length; Normalize the numerical data to ensure that all features are on the same scale; divide the preprocessed dataset into training set, validation set and test set; Risk assessment module, used to build risk assessment models, model evaluation and adjustment; The monitoring and early warning module includes a monitoring unit, a judgment unit and an early warning unit. The monitoring unit is used to monitor the transaction behaviors on the e-commerce platform in real time, input new transaction behaviors into the risk assessment module, and obtain the predicted probability; the judgment unit judges whether the obtained predicted probability is higher than the threshold. Transaction behaviors higher than the threshold are abnormal behaviors, and transaction behaviors lower than the threshold are normal behaviors: the early warning unit, abnormal behaviors are used by the system to automatically trigger the early warning mechanism, and notify users and platform administrators through emails, SMS, APP push, etc.; The transaction verification module is used to introduce biometric technology and secondary verification methods in the transaction process to ensure the authenticity and legality of the transaction. The verification mechanism is combined with the risk assessment model to form a double protection network to enhance transaction security; The fraud analysis module includes an analysis unit and a tracing unit. The analysis unit conducts in-depth analysis on confirmed fraudulent behaviors, extracts fraud characteristics, optimizes risk assessment models, establishes a fraud behavior database, and records the behavior patterns and characteristics of fraudsters. The tracing unit provides fraud tracing functions to help the platform quickly locate the source of fraud and take corresponding measures.

Images