Optimised Prediction Model for Stock Market Trend Analysis

The main objective of this work is to add to the academic understanding of stock market analysis using some well defined algorithms and machine learning techniques. Stock price forecasting is a popular and important topic in financial studies and at academic levels. Share Market is not a neat place for analyzing since there are no significant rules to estimate or predict the price of share in the share market. Many a method like technical analysis, fundamental analysis, time series analysis and statistical analysis, etc. have been used in an attempt to analyze the share trends in the market but none of these methods have so far proved to be a universal approach for acceptance as a prediction tool. The intricacy while analyzing market trends is that they have a dependency on a number of external factors some of which are not under one's control. The goal of this work is to analyze stock market trends using some machine learning and nature inspired techniques, these were first studied and then implemented (a few of them used in this paper are Decision Tree, PSO, Black-Hole, Naïve Bayes.) After analyzing the trends with the help of standard techniques, we then proposed an entirely new approach to analyze stock market indices over which accuracy is calculated and compared over different techniques and algorithms. We outline the design of the proposed model with its salient features and customizable parameters. We finally tested our model on the one year of Nifty stock index dataset at real time where we analyzed the values on the basis of data from the past days for three months.