Multi-dimensional distributed abnormal transaction behavior detection method

Abstract

The invention discloses a multi-dimensional distributed abnormal transaction behavior detection method. The method comprises the following steps: firstly, mining multi-dimensional original feature examples in network transaction behaviors before and in transactions; secondly, providing an automatic feature learning and fusion algorithm MSDAE based on deep learning to remove redundancy and noise inoriginal features and automatically learn implicit and representative features; and finally, a parallel distributed integrated framework SpaEnsemble based on Apache Spark is provided to realize efficient and rapid analysis and detection of large-scale abnormal transaction behaviors. The method has a wide application prospect in the field of network security.

Description

Technical field [0001] The invention belongs to the field of network security, and specifically relates to a multi-dimensional distributed abnormal transaction behavior detection method. Background technique [0002] Blockchain is a new decentralized basic framework and distributed computing paradigm that has gradually emerged with the development of digital cryptocurrencies such as Bitcoin. It uses an orderly chained data structure to store data and uses consensus algorithms to update data. , The use of cryptography technology to ensure its data security, etc., has the characteristics of non-tampering, decentralization, trustlessness, traceability, collective maintenance and security. Among them, Blockchain 2.0 has been used in the fields of finance, logistics, energy, medicine and health, and its biggest feature is the introduction of smart contracts, which can enable developers to implement complex blockchain applications due to its Turing completeness. . Although smart cont...

AI Technical Summary

Problems solved by technology

[0003] At present, the detection of abnormal transaction behaviors in this technical field mainly has the following problems: (1) single source of features makes it difficult to reasonably represent transaction behaviors; (2) too much reliance on manual or prior knowledge in the feature selection process leads to high cost but expandable (3) In the process of classification of trading behaviors, the performance of traditional classification algorithms depends on the distribution of data in the training set, such as the balance of positive and negative samples. The accuracy rate is low; (4) At present, most of the research is based on small data samples, and it is difficult to meet the characteristics of large data samples, such as massive, multi-dimensional, high-speed and changeable, complex internal correlations, and high real-time requirements for abnormal transaction behavior detection.

Images