Stock Market Prediction Using Clustering with Meta-Heuristic Approaches

Various examinations are performed to predict the stock values, yet not many points at assessing the predictability of the direction of stock index movement. Stock market prediction with data mining method is a standout amongst the most paramount issues to be researched and it is one of the interesting issues of stock market research over several decades. The approach of advanced data mining tools and refined database innovations has empowered specialists to handle the immense measure of data created by the dynamic stock market. Data mining strategies have been utilized to reveal hidden patterns and predict future patterns and practices in financial markets to help financial investors make qualitative choice. In this paper, the consistency of stock index movement of the well-known Indian Stock Market indices NSE-NIFTY are examined with the assistance of famous data mining strategies known as Clustering. Clustering is the methodology of grouping the alike indices into clusters. It likewise audits three of the meta-heuristics clustering algorithms: PSO-K-Means, Bat Algorithm, and firefly Algorithm. These strategies are implemented and tested against a Stock Market Index Movement Dataset. The performance of the aforementioned procedures is compared based on "integrity of clustering " assessment measures. The investigation is used to the NSE-NIFTY and BSE-NIFTY for the period from January 2011 to April 2014