Detection of Stock Market Manipulation Using Deep Learning

Abstract

Abstract The stock market is a vast trading environment that handles millions of transactions, making it challenging for regulatory bodies to identify fraudulent activities manually. However, unsupervised deep learning techniques provide a promising solution for detecting market manipulation. Inflating or deflating stock prices to gain an advantage is a form of market manipulation. We propose to examine how stock market manipulation can be detected using market structure analysis in the paper. The gathering of data involved utilizing information obtained from the websites of National Stock Exchange (NSE) and the Bombay Stock Exchange (BSE). The efficacy of an approach utilizing generative models is particularly noteworthy in this domain, showing significant potential for quickly identifying specific local irregularities in the data for anomaly detection and potential recognition of market manipulation. This efficiency addresses a common challenge faced by deep learning approaches. By exploring this matter, we aim to contribute insightful perspectives that could aid both investors and regulatory bodies in effectively understanding and managing the risks associated with stock market manipulation.