Abnormal key account discovery method and system based on graph neural network, equipment and storage medium
A neural network and discovery method technology, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve the problems of inaccurate analysis results and low usability, and achieve broad application prospects, time saving, and increased accuracy. Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0066] An abnormal key account discovery method based on graph neural network, such as figure 1 , Figure 5 shown, including the following steps:
[0067] (1) Data preprocessing: perform operations such as data cleaning, key data item extraction, and account transaction relationship construction within the organization on the historical transaction records of abnormal financial accounts;
[0068] Data cleaning refers to cleaning all transaction data related to normal accounts, and only retaining historical transaction records where both parties to the transaction are abnormal financial accounts;
[0069] The present invention mainly focuses on the financial transaction network within the abnormal organization, and the purpose is to discover the key accounts of the abnormal organization from the transaction behavior between the financial accounts within the abnormal organization, including the high-level accounts of the organization (organizers, leaders) As well as the purcha...
Embodiment 2
[0084] According to a method for discovering abnormal key accounts based on a graph neural network described in Embodiment 1, the difference is that:
[0085] This method is applicable to computing platforms with a CPU version or performance not lower than intel i5, a memory of more than 4G, and a Linux operating system. The Linux system needs to be configured with the Tensorflow and Keras frameworks. In the above configuration, a computing platform with powerful GPU computing capability is a better choice for running this method.
[0086] In step (1), the threshold method is used for data cleaning, which specifically means: if the absolute difference of the number of capital inflows and outflows in the current transaction record is less than a given threshold, it is considered a normal account and cleaned, otherwise, it is retained.
[0087] The present invention adopts the threshold method to clean the data. There is a big difference in the distribution of capital inflows and...
Embodiment 3
[0123] An abnormal key account discovery system based on a graph neural network, used to realize the abnormal key account discovery method based on a graph neural network described in Embodiment 1 or 2, including a data preprocessing module, a financial transaction network graph of an abnormal organization Building blocks, key account discovery modules for abnormal organizations;
[0124] The historical transaction records of abnormal financial accounts will obtain the transaction relationship of accounts within the organization after preprocessing such as data cleaning. The data preprocessing module is used to: sequentially perform operations such as data cleaning, key data item extraction, and account transaction relationship construction within the organization on historical transaction records of abnormal financial accounts; the abnormal organization financial transaction network diagram construction module is used to : According to the intra-organizational account transac...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More