Prediction of Trading Malpractices / Patterns using fuzzy logic

Abstract

Someone can argue that the most dynamic places in today’s world are the stock exchanges and this is quite obvious with concepts such as High Frequency Trading and Sponsored Access. As a result of these concepts the stock exchanges are getting flooded with high order rates and volumes. Therefore surveilling market places has become even more challenging and the current systems are struggling to cope up with the trades and the data mess. The regulators are however pushing exchanges to regulate their order flows and to detect manipulative behavior. To be fair on them this is prudent given the fact that stock market crashes we faced in the recent past. Stock market crash happened in U.S. in the last year is a prime example for this. Therefore having this in mind the author hereby presents a novel approach to market monitoring using subtractive clustering based fuzzy system identification method which can cope up even at high order rates. Artificial intelligence was extensively used in order to predict trading malpractices via fuzzy logic which to identify complex manipulations and behaviors. The proposed system is proven to capture all the manipulative behaviors including edge conditions, in order to maintain a fair and efficient market and to support users with efficient manipulative scenarios. The study discussed and addressed broad issues of fuzzy logic related to Market Surveillance, and there exist many areas and opportunities for further studies in this area. One such work is to embed the capability of handling text data type to cater for Market announcements and Rumors which would further help Surveillance users to discover new findings. Also study can be elaborated to identify adaptive neuro-fuzzy inference system to identify optimal values for fuzzy sets rather than depending on expert knowledge as done in this study. ii