Forecasting stock market prices using mixed ARIMA model: a case study of Indian pharmaceutical companies

Abstract

Many investors in order to predict stock prices use various techniques like fundamental analysis and technical analysis and sometimes rely on the discussions provided by various stock market analysts. ARIMA is a part of time-series analysis under prediction algorithms, and this paper attempts to predict the share prices of selected pharmaceutical companies in India, listed under NIFTY100, using the ARIMA model. A sample size of 782 time-series observations from January 1, 2017 to December 31, 2019 for each selected pharmaceutical firm has been considered to frame the ARIMA model. ADF test is used to verify whether the data are stationary or not. For ARIMA model estimation, significant spikes in the correlogram of ACF and PACF have been observed, and many models have been framed taking different AR and MA terms for each selected company. After that, 5 best models have been selected, and necessary inculcation of various AR and MA terms has been made to adjust the models and choose the best adjusted ARIMA model for each firm based on Volatility, adjusted R-squared, and Akaike Information Criterion. The results could be used to analyze the stock prices and their prediction in-depth in future research efforts.