Limitation of ARIMA in extremely collapsed market: A proposed method

The prediction of equity market is a perplexing task because of puzzling nature of the stock price. A large variety of factors influence the price of stocks that causes the investors in trouble to predict the nature of stock. Researchers proposed various methods to forecast the upcoming price of stocks by figuring out the nature of stock and by computing internal and external factors. Auto Regressive Integrated Moving Average (ARIMA) Box-Jekins method is one of the eminent methods that forecast the future value of stock based on previous time series data. But while experimenting, we have found a crack point where ARIMA showed unstable behavior i.e. it returned 3 types of values (fixed, negative and positive) in prediction. This was after the prices of stock extremely collapsed. To solve this problem, we have proposed a new model that also predicts the future prices from previous prices but obtained greater accuracy than the ARIMA and also solve the negative and fixed value prediction problem occurred in ARIMA.