A Novel Hybrid Method for Generating Association Rules for Stock Market Data

Abstract – The aim of the current research is to extract the knowledge from stock market data to help investors make more profit. We have used NSE (National Stock Exchange) historical data for six years. We have also pre-processed stock market data. We have used hybrid method which consists of two widely used algorithms FP-growth to find frequent patterns and Discovery rule algorithm proposed by Agrawal’94 to get association rules for two different steps of association rule mining. The main focus in our research has been the accuracy of the rules. The goal of the research is to find dependencies among different stock companies in the stock market and generate rules from inter-day transactions that would benefit stock market traders.

Keywords: Stock market, NSE, Association Rule Mining, FPgrowth

I. INTRODUCTION

Now days, utilisation of large amount of data from stored databases and to analyse it is one popular trend prevalent amongst researchers. Extraction of hidden patterns or corelation among related attributes from stock market analysis is one of the major research areas. The stock market is the place where investors purchase stocks become shareholders for financial achievements of the companies. If the price of stocks goes down then the investor losses money and if stock prices goes up the investor makes profit. So, Investors trade on shares for making profit. It is desirable to generate interesting rules or patterns to help the investor make profit. Stock Market Analysis and prediction is done by lots of researchers to extract associated and co-related rules or patterns to increase the profit in stock market. Stock market data is available in different forms like stock indices [7], interday-transaction [8][9][17][18][4], intra-day-transaction [11] and many more forms. Stock market has the number of product types for trading like equity, mutual funds, derivatives etc. Since last many years, researchers have taken keen interest and worked in this area. The stock market fluctuations and unpredictable changes require discovering some concrete and interesting rules from it, so that investors can be guided to make safe decisions for investment to get maximum profit.
Read More