An Intelligent Technique for Stock Market Prediction

Abstract

Economy of a country greatly depends on the stock market sector. People may gain profit by proper investments in stock markets or may lose their entire life savings by wrong investments. Previously, the world has witnessed many stock markets disaster, crippling economic condition of an entire nation. Bangladesh also faced similar situation in the near past. This can be avoided if the market can be predicted in advance. By this prediction, a broker may be able to find some irregularities which may alert them in advance. Also, the stock market prediction can be a useful tool for the market regulatory committee. This prediction may help them to take necessary steps to avoid potentially harmful transactions. In this paper, we propose a model for stock market prediction based on large amount of historical data and machine learning approach. For the experimental purpose, we have collected data from two of the Bangladeshi stock exchanges: Dhaka Stock Exchange (DSE) and Chittagong Stock Exchange (CSE). To perform this task, we have applied web crawling to crawl data from source web sites and applied data parsing to get desired data for training our system. Finally, linear regression approach of machine learning was used to get the prediction for individual stock securities.