Stock Price Volatility and Forecasting using Neural Network based ARIMA Model

Abstract

Trading in stock markets is not an easy task and requires expertise and knowledge that improve and increase the chances of making more profits and ensuring that you make profitable decisions at all times. Financial Experts and data analysts make use of various algorithms and neural networks to predict the value of stock depending upon the available information and then use the outcomes to analyze and make their trading decision.

In this concept the research was conducted by taking top sectors from the NSE/BSE. Each sector comprises of five companies based on the weightage value. The study was conducted with different sectors like, IT sector, Automobile sector, Banking sector, Pharmaceutical sector and FMCG sector. All the sectors taken for the study is highly volatile compared to other sectors in NSE/BSE. Hence it is very essential to study on the nexus between the Indian Stock Market and selected companies behaviour. The research has been segregated into four segments based on the objectives framed. To ascertain the stock price volatility the daily return was used. To identify the risk and relationship the OLS model was used. Finally the ARIMA model was build to find the impact of today’s stock price on its historical stock price for the selected stocks. Based on the model fit forecasting was done for the short period of three months.