Cross-industry anti-fraud early warning method and system based on big data

Abstract

The invention relates to the technical field of anti-fraud early warning, and discloses a cross-industry anti-fraud early warning method and system based on big data, and the method comprises the following steps: collecting first comprehensive information and logistics information of n continuous time points including a current time point and a previous time point; generating first two-dimensional structure data based on the first comprehensive information, and generating second two-dimensional structure data based on the logistics information; inputting the collected first comprehensive information into a risk assessment model, and outputting a risk assessment report; and the judgment result is decoded and notified to consumers and related supervision departments. The system has the beneficial effects that the logistics information of each logistics company of a plurality of countries is re-integrated, and then anti-fraud early warning is performed through the integrated information, so that the problems of data diversity and consistency caused by multi-channel logistics tracking and transportation delay are solved, and real-time monitoring and anomaly detection of the logistics state are ensured.

Description

Technical Field

[0001] The present invention relates to the field of anti-fraud early warning technology, and more specifically, to a cross-industry anti-fraud early warning method and system based on big data. Background Art

[0002] In today's era of globalization, cross-border e-commerce and logistics industries are booming. However, with the rapid expansion of the industry, cross-border fraud is becoming increasingly serious, becoming a major problem that hinders the healthy development of the industry.

[0003] The reputation of e-commerce platforms and logistics companies is the foundation of their survival and development. The occurrence of cross-border fraud incidents will seriously damage the reputation of enterprises and reduce consumers' trust in enterprises. Conducting cross-industry anti-fraud warnings can help enterprises detect and deal with fraud in a timely manner, reduce the occurrence of fraud incidents, and thus maintain the reputation of enterprises.

[0004] Since cross-border logistics involves transportation in multiple countries and multiple logistics service providers, the diversity and consistency of data sources are increased, affecting the real-time monitoring and anomaly detection of logistics status. Summary of the invention

[0005] The present invention provides a cross-industry anti-fraud early warning method and system based on big data, which solves the technical problem in related technologies that cross-border logistics involves transportation in multiple countries and multiple logistics service providers, which increases the diversity and consistency of data sources and affects the real-time monitoring of logistics status and anomaly detection.

[0006] The present invention provides a cross-industry fraud prevention early warning method based on big data, comprising: Step 100, collecting first comprehensive information and logistics information of a total of n consecutive time points including the current time point and previous time points, the first comprehensive information including: cross-border e-commerce transaction data, external data sources; Step 200, generating first two-dimensional structure data based on the first comprehensive information, and generating second two-dimensional structure data based on the logistics information; Step 300, input the collected first comprehensive information into the risk assessment model, the risk assessment model includes a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a first feature fusion layer, a first output layer, and a second output layer, wherein the first intermediate layer is used to input the first two-dimensional structure data, output the first intermediate representation data to the second intermediate layer, the second intermediate layer outputs the second intermediate representation data to the first feature fusion layer, the third intermediate layer is used to input the second two-dimensional structure data, output the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer and the first feature fusion layer, the second output layer outputs the integrated logistics information, the first feature fusion layer outputs the first fusion representation data to the first output layer, and the first output layer outputs the risk assessment report; Step 400, decode the determination result and notify the consumer and relevant regulatory authorities.

[0007] In a preferred embodiment, cross-border e-commerce transaction information includes: order number, product category, picture, quantity, amount, user ID, registration information, purchase history, payment method, payment time, and payment status; External data information includes: known fraudulent users, suspicious IP addresses, consumer complaint records, user reviews, and public opinion analysis; Logistics information includes: order number, shipment time, arrival time, logistics status, place of shipment, destination, transshipment center, transportation route, estimated transportation time, actual transportation time, reason for delay, logistics company name, carrier information, logistics company contact information, package number, cargo type, weight and size.

[0008] In a preferred embodiment, the No. 1 two-dimensional structure data includes a No. 1 data matrix and a No. 1 relationship matrix, and a unit of the No. 1 data matrix represents a No. 1 one-dimensional structure data of a No. 1 independent object, and the No. 1 independent object includes an e-commerce platform, a logistics company, a commodity, and a user, and a unit of the No. 1 data matrix only includes the first comprehensive information of the No. 1 independent object represented by it; The element in the i-th row and j-th column of the No. 1 relationship matrix represents the association between the i-th unit of the No. 1 data matrix and the No. 1 independent object represented by the j-th unit. If there is an association, the value of this element of the No. 1 relationship matrix is ​​1, otherwise it is 0; The relationship between e-commerce platforms and products means that product information will be displayed on the e-commerce platform; The connection between users and e-commerce platforms means that users’ purchase information will be displayed on the e-commerce platform; The association between a product and a user means that: the product is a product purchased by the user; The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the first comprehensive information collected at the t-th moment.

[0009] In a preferred embodiment, the second two-dimensional structure data includes a second data matrix and a second relationship matrix, and one unit of the second data matrix represents the second one-dimensional structure data of a second independent object, the second independent object includes a logistics company and a commodity, and one unit of the second data matrix only contains the logistics information of the second independent object represented by it; The element in the i-th row and j-th column of the second relationship matrix represents the association between the i-th unit of the second data matrix and the second independent object represented by the j-th unit. If there is an association, the value of this element of the second relationship matrix is ​​1, otherwise it is 0; The connection between logistics companies and goods means that the circulation information of goods will be displayed on the logistics company's platform; The second one-dimensional structure data includes n data items sorted by time, and the tth data item represents the logistics information collected at the tth moment.

[0010] In a preferred embodiment, the risk assessment report includes whether it is a risky transaction, the risky transaction level, and the handling method.

[0011] In a preferred embodiment, the integrated logistics information includes the integrated logistics information of various logistics companies.

[0012] In a preferred embodiment, the first comprehensive information and logistics information are input into the model after feature engineering; First, the comprehensive information and logistics information include text mode and image mode. For the data of various modes, the corresponding feature engineering method is used to encode and generate features, and then splice them; The information of text modality is used as a feature engineering method through the Word Embedding algorithm; The information of image modality is used as a feature engineering method through CNN algorithm.

[0013] In a preferred embodiment, the loss function during the training of the risk assessment model is a cross entropy loss function, and the calculation formula is as follows: ; in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. The risk assessment model predicts The training samples belong to The probability of the classification category, Represents the logarithmic function with the natural constant e as the base.

[0014] In a preferred embodiment, the third intermediate layer inputs the second two-dimensional structure data and outputs the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer, and the second output layer outputs the integrated logistics information. This process is pre-trained separately, and the training loss function is a combination of the cross entropy loss function and the mean square error loss function. The calculation formula is as follows: ; in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. It is the predicted The training samples belong to The probability of the classification category, Represents the logarithmic function with the natural constant e as the base.

[0015] ; ; in represents the loss value, represents the total number of training samples, Indicates The actual measured value or parameter of the samples, The output of the model is The measured values ​​or parameters of samples, represents the total loss value, and are the first and second weight coefficients respectively.

[0016] The beneficial effects of the present invention are as follows: the present invention re-integrates the logistics information of various logistics companies in multiple countries, and then uses the integrated information for anti-fraud warning, thereby solving the data diversity and consistency problems caused by multi-channel logistics tracking and transportation delays, and ensuring real-time monitoring and anomaly detection of logistics status. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a cross-industry anti-fraud early warning method based on big data of the present invention. DETAILED DESCRIPTION

[0018] The subject matter described herein will now be discussed with reference to example implementations. It should be understood that the discussion of these implementations is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the present specification. Various examples may omit, replace, or add various processes or components as needed. In addition, the features described in some examples may also be combined in other examples.

[0019] At least one embodiment of the present invention discloses a cross-industry fraud prevention and early warning method based on big data, such as Figure 1 As shown, including: Step 100, collecting first comprehensive information and logistics information of a total of n consecutive time points including the current time point and previous time points, the first comprehensive information including: cross-border e-commerce transaction data, external data sources; Cross-border e-commerce transaction information includes: order number, product category, picture, quantity, amount, user ID, registration information, purchase history, payment method, payment time, and payment status; External data information includes: known fraudulent users, suspicious IP addresses, consumer complaint records, user reviews, and public opinion analysis; Logistics information includes: order number, shipping time, arrival time, logistics status, shipping place, destination, transshipment center, transportation route, estimated transportation time, actual transportation time, delay reason, logistics company name, carrier information, logistics company contact information, package number, cargo type, weight and size; In one embodiment of the present invention, information of different companies is different, and some information may not exist. The non-existent information is marked as a null value.

[0020] Step 200, generating first two-dimensional structure data based on the first comprehensive information, and generating second two-dimensional structure data based on the logistics information; The No. 1 two-dimensional structure data includes a No. 1 data matrix and a No. 1 relationship matrix. A unit of the No. 1 data matrix represents a No. 1 one-dimensional structure data of a No. 1 independent object. The No. 1 independent object includes an e-commerce platform, a logistics company, a commodity, and a user. A unit of the No. 1 data matrix only includes the first comprehensive information of the No. 1 independent object it represents. The element in the i-th row and j-th column of the No. 1 relationship matrix represents the association between the i-th unit of the No. 1 data matrix and the No. 1 independent object represented by the j-th unit. If there is an association, the value of this element of the No. 1 relationship matrix is ​​1, otherwise it is 0; The relationship between e-commerce platforms and products means that product information will be displayed on the e-commerce platform; The connection between users and e-commerce platforms means that users’ purchase information will be displayed on the e-commerce platform; The association between a product and a user means that: the product is a product purchased by the user; The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the first comprehensive information collected at the t-th moment; The second two-dimensional structure data includes a second data matrix and a second relationship matrix. One unit of the second data matrix represents the second one-dimensional structure data of a second independent object. The second independent object includes a logistics company and a commodity. One unit of the second data matrix only contains the logistics information of the second independent object it represents. The element in the i-th row and j-th column of the second relationship matrix represents the association between the i-th unit of the second data matrix and the second independent object represented by the j-th unit. If there is an association, the value of this element of the second relationship matrix is ​​1, otherwise it is 0; The connection between logistics companies and goods means that the circulation information of goods will be displayed on the logistics company's platform; The second one-dimensional structure data includes n data items sorted by time, and the tth data item represents the logistics information collected at the tth moment; Step 300, input the collected first comprehensive information into the risk assessment model, the risk assessment model includes a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a first feature fusion layer, a first output layer, and a second output layer, wherein the first intermediate layer is used to input the first two-dimensional structure data, output the first intermediate representation data to the second intermediate layer, the second intermediate layer outputs the second intermediate representation data to the first feature fusion layer, the third intermediate layer is used to input the second two-dimensional structure data, output the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer and the first feature fusion layer, the second output layer outputs the integrated logistics information, the first feature fusion layer outputs the first fusion representation data to the first output layer, and the first output layer outputs the risk assessment report; In one embodiment of the present invention, the calculation formula of the first intermediate layer is as follows: ; ; ; ; in, , , represents the first, second, and third weight parameters, , , represents the first, second and third bias parameters, represents the dot product, , and They represent the first, second and third intermediate states of the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. represents the t-th data item of the one-dimensional structure data corresponding to the v-th unit of the data matrix in the two-dimensional structure data. and They represent the t-th and t-1-th first intermediate representation data of the v-th unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. n≥t≥1, n represents the total number of data items of the one-dimensional structure data No. 1, and when t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.

[0021] In one embodiment of the present invention, the calculation formula of the second intermediate layer is as follows: ; in The second intermediate representation data of the vth unit of the data matrix No. 1, is the set of cells of data matrix No. 1 that are associated with the v-th cell of data matrix No. 1, , , and They represent the t-th first intermediate representation data of the v-th unit and the u-th unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. is a learnable scalar parameter, represents the two-dimensional recognition weight parameter, and MLP represents the multi-layer perceptron.

[0022] In one embodiment of the present invention, the calculation formula of the third intermediate layer is as follows: The long short-term memory network includes a forget gate, an input gate, a cell state, and an output gate. The output of the forget gate of the vth unit of the second data matrix in the second two-dimensional structure data is for: ; in, is the weight matrix of the forget gate, is the bias term of the forget gate, It is the t-1th intermediate representation data of the vth unit of the second data matrix in the second two-dimensional structure data. and The splicing, is the sigmoid function; represents the t-th data item of the second one-dimensional structure data corresponding to the v-th unit of the second data matrix in the second two-dimensional structure data; The tth output of the input gate of the vth unit of the second data matrix in the second two-dimensional structure data The t-th candidate cell state of the v-th unit of the second data matrix in the second two-dimensional structure data The calculation method is: ; ; in, is the weight matrix of the input gate, is the bias term of the input gate, is the hyperbolic tangent activation function, is the weight matrix for calculating candidate cell states, is the bias term of the candidate cell state; The t-th cell state of the v-th unit of the data matrix No. 2 in the two-dimensional structure data No. 2 is the t-1th cell state of the vth unit of the second data matrix in the second two-dimensional structure data and Weighted synthesis, the calculation method is: ; in, Represents element-by-element multiplication operation; The tth activation value of the output gate of the vth unit of the second data matrix in the second two-dimensional structure data and the t-th third intermediate representation data of the v-th unit of the second data matrix in the second two-dimensional structure data The calculation method is: ; ; in, is the weight matrix of the output gate, is the bias term of the output gate.

[0023] In one embodiment of the present invention, the calculation formula of the fourth intermediate layer is as follows: ; ; ; in The fourth intermediate representation data of the vth unit of the second data matrix, , , and They represent the t-th third intermediate representation data of the v-th unit and the u-th unit of the second data matrix in the second two-dimensional structure data, respectively. represents the attention coefficient of the vth and uth units, represents the attention coefficient of the vth and sth units, represents the normalized attention coefficients of the vth and uth units, represents the set of cells of the second data matrix that are associated with the vth cell of the second data matrix, represents an exponential function with a natural constant as base, represents the trainable attention vector parameter, CONCAT represents the concatenation operation, represents transpose, represents the two-dimensional recognition weight parameter, and LeakyReLU is the LeakyReLU activation function.

[0024] In one embodiment of the present invention, the calculation formula of the first feature fusion layer is as follows: ; in, represents the first fusion state, represents the concatenation function, represents the sum weight matrix, represents the summation bias parameter, The second intermediate representation data of the vth unit of the data matrix No. 1, is the set of cells of data matrix No. 1 that are associated with the v-th cell of data matrix No. 1, The fourth intermediate representation data of the vth unit of the second data matrix, Represents the set of cells of the second data matrix that are associated with the vth cell of the second data matrix.

[0025] In one embodiment of the present invention, the calculation formula of the first output layer is as follows: ; in Represents a vector, the components of which represent whether it is a risky transaction, the risky transaction level, and the handling method. represents the first fusion state, is the first weight parameter, is the first bias parameter, Represents the sigmoid function.

[0026] In one embodiment of the present invention, whether it is a risky transaction is divided into yes and no, corresponding to two values ​​of 1 and 0 respectively.

[0027] In one embodiment of the present invention, the risk levels are divided into four levels: normal, low, medium, and high.

[0028] In one embodiment of the present invention, different risk levels correspond to different processing methods. Normally, no processing method is adopted. Low-risk ones are recorded by conventional monitoring and systems. Medium-risk ones are further reviewed manually. Further communication and more detailed information are conducted on the transaction to identify fraud. For high-risk ones, transactions are suspended, and transaction funds are urgently frozen. Funds and goods are controlled and managed, and in-depth investigations are conducted.

[0029] In one embodiment of the present invention, the calculation formula of the second output layer is as follows: ; in represents a vector, the i-th component of the vector represents the integrated logistics information of the i-th logistics company, The fourth intermediate representation data of the vth unit of the second data matrix, represents the set of cells of the second data matrix that are associated with the vth cell of the second data matrix, is the second weight parameter, is the second bias parameter, Represents the sigmoid function.

[0030] In one embodiment of the present invention, the first comprehensive information and logistics information are input into the model after feature engineering; First, the comprehensive information and logistics information include text mode and image mode. For the data of various modes, the corresponding feature engineering method is used to encode and generate features, and then splice them; The information of text modality is used as a feature engineering method through the Word Embedding algorithm; The information of image modality is used as a feature engineering method through CNN algorithm; Step 400, decode the determination result and notify the consumer and relevant regulatory authorities.

[0031] In one embodiment of the present invention, the loss function during the training of the risk assessment model is a cross entropy loss function, and the calculation formula is as follows: ; in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. The risk assessment model predicts The training samples belong to The probability of the classification category, Represents the logarithmic function with the natural constant e as the base.

[0032] In one embodiment of the present invention, whether it is a risky transaction, the risky transaction level, and the processing method are freely combined to form a classification category.

[0033] In one embodiment of the present invention, the third intermediate layer inputs the second two-dimensional structure data and outputs the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer, and the second output layer outputs the integrated logistics information. This process is pre-trained separately, and the training loss function is a combination of the cross entropy loss function and the mean square error loss function. The calculation formula is as follows: ; in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. It is the predicted The training samples belong to The probability of the classification category, Represents the logarithmic function with the natural constant e as the base.

[0034] ; ; in represents the loss value, represents the total number of training samples, Indicates The actual measured value or parameter of the samples, The output of the model is The measured values ​​or parameters of samples, represents the total loss value, and are the first and second weight coefficients respectively.

[0035] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation mode. The above-mentioned specific implementation mode is merely illustrative and not restrictive. Under the guidance of this embodiment, ordinary technicians in this field can also make more forms of equivalent embodiments, all of which are within the protection of this embodiment.

Claims

1. A cross-industry anti-fraud early warning method based on big data, characterized in that: The following steps are involved: Step 100, collecting first comprehensive information and logistics information of a total of n consecutive time points including the current time point and previous time points, the first comprehensive information including: cross-border e-commerce transaction data, external data sources; Step 200, generating first two-dimensional structure data based on the first comprehensive information, and generating second two-dimensional structure data based on the logistics information; Step 300, input the collected first comprehensive information into the risk assessment model, the risk assessment model includes a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a first feature fusion layer, a first output layer, and a second output layer, wherein the first intermediate layer is used to input the first two-dimensional structure data, output the first intermediate representation data to the second intermediate layer, the second intermediate layer outputs the second intermediate representation data to the first feature fusion layer, the third intermediate layer is used to input the second two-dimensional structure data, output the third intermediate representation data to the fourth intermediate layer, the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer and the first feature fusion layer, the second output layer outputs the integrated logistics information, the first feature fusion layer outputs the first fusion representation data to the first output layer, and the first output layer outputs the risk assessment report; Step 400, decode the determination result and notify the consumer and relevant regulatory authorities.

2. According to the big data-based cross-industry fraud prevention early warning method of claim 1, it is characterized in that: Cross-border e-commerce transaction information includes: order number, product category, picture, quantity, amount, user ID, registration information, purchase history, payment method, payment time, and payment status; External data information includes: known fraudulent users, suspicious IP addresses, consumer complaint records, user reviews, and public opinion analysis; Logistics information includes: order number, shipment time, arrival time, logistics status, place of shipment, destination, transshipment center, transportation route, estimated transportation time, actual transportation time, reason for delay, logistics company name, carrier information, logistics company contact information, package number, cargo type, weight and size.

3. According to the big data-based cross-industry fraud prevention early warning method of claim 1, it is characterized in that: The No. 1 two-dimensional structure data includes a No. 1 data matrix and a No. 1 relationship matrix. A unit of the No. 1 data matrix represents a No. 1 one-dimensional structure data of a No. 1 independent object. The No. 1 independent object includes an e-commerce platform, a logistics company, a commodity, and a user. A unit of the No. 1 data matrix only includes the first comprehensive information of the No. 1 independent object it represents. The element in the i-th row and j-th column of the No. 1 relationship matrix represents the association between the i-th unit of the No. 1 data matrix and the No. 1 independent object represented by the j-th unit. If there is an association, the value of this element of the No. 1 relationship matrix is ​​1, otherwise it is 0; The relationship between e-commerce platforms and products means that product information will be displayed on the e-commerce platform; The connection between users and e-commerce platforms means that users’ purchase information will be displayed on the e-commerce platform; The association between a product and a user means that: the product is a product purchased by the user; The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the first comprehensive information collected at the t-th moment.

4. According to the big data-based cross-industry fraud prevention early warning method of claim 1, it is characterized in that: The second two-dimensional structure data includes a second data matrix and a second relationship matrix. One unit of the second data matrix represents the second one-dimensional structure data of a second independent object. The second independent object includes a logistics company and a commodity. One unit of the second data matrix only contains the logistics information of the second independent object it represents. The element in the i-th row and j-th column of the second relationship matrix represents the association between the i-th unit of the second data matrix and the second independent object represented by the j-th unit. If there is an association, the value of this element of the second relationship matrix is ​​1, otherwise it is 0; The connection between logistics companies and goods means that the circulation information of goods will be displayed on the logistics company's platform; The second one-dimensional structure data includes n data items sorted by time, and the tth data item represents the logistics information collected at the tth moment.

5. According to the big data-based cross-industry fraud prevention early warning method of claim 1, it is characterized in that: The risk assessment report includes whether it is a risky transaction, the risk transaction level, and the handling method.

6. The cross-industry anti-fraud early warning method based on big data according to claim 1 is characterized in that: The integrated logistics information includes the integrated logistics information of each logistics company.

7. The cross-industry fraud prevention early warning method based on big data according to claim 1 is characterized in that: First, comprehensive information and logistics information are input into the model after feature engineering; First, the comprehensive information and logistics information include text mode and image mode. For the data of various modes, the corresponding feature engineering method is used to encode and generate features, and then splice them; The information of text modality is used as a feature engineering method through the Word Embedding algorithm; The information of image modality is used as a feature engineering method through CNN algorithm.

8. The cross-industry fraud prevention early warning method based on big data according to claim 1 is characterized in that: The loss function during the training of the risk assessment model is the cross entropy loss function, and the calculation formula is as follows: in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. The risk assessment model predicts The training samples belong to The probability of the classification category, Represents the logarithmic function with the natural constant e as the base.

9. The cross-industry fraud prevention early warning method based on big data according to claim 1 is characterized in that: The third intermediate layer inputs the second two-dimensional structure data and outputs the third intermediate representation data to the fourth intermediate layer, and the fourth intermediate layer outputs the fourth intermediate representation data to the second output layer. The second output layer outputs the integrated logistics information. This process is pre-trained separately. The training loss function is a combination of the cross entropy loss function and the mean square error loss function. The calculation formula is as follows: ； in represents the loss value, represents the total number of classification categories, represents the total number of training samples, It is The symbol value of the training sample, if the The true classification of the training samples is The value of a classification category is 1, otherwise it is 0. It is the predicted The training samples belong to The probability of the classification category, It represents the logarithmic function with the natural constant e as the base; ； ； in represents the loss value, represents the total number of training samples, Indicates The actual measured value or parameter of the samples, The output of the model is The measured values ​​or parameters of samples, represents the total loss value, and are the first and second weight coefficients respectively.

10. A cross-industry anti-fraud early warning system based on big data, characterized in that: It is used to execute a cross-industry anti-fraud early warning method based on big data as described in any one of claims 1-9.

Images