Forecasting the stock market using ARIMA modeling and foresight

Abstract

In this work, the future prices of the sample companies are predicted using the Box and Jenkins-popularized ARIMA model. The process of methodically discovering, estimating, diagnosing, and forecasting time series is known as ARIMA model construction, and it is driven by empirical evidence. Only stationary time series are suitable for use with the ARIMA model. Because the time series pertaining to the stock market are non-stationary, the underlying stochastic process's properties fluctuate with time. An ARIMA model's AR and MA components only apply to time series that are stationary, in order to forecast the stock market and take into account the stationary time series in this research. Each sample company's model selection must have a foundation, which is demonstrated by determining the ACF, PACF, AIC, and SBC values. The ACF and PACF error plots, which are shown in the images, are crucial. The forecast graph visualization using sample data, which is provided in this chapter for each organization, will be excellent.