Hybrid time series and artificial neural network models for forecasting of the banking stock prices during Covid-19 pandemic

Abstract

The stocks are one of a variety of securities that are traded in general through the stock exchange. One sector that is quite large in the Indonesian capital market is banking stocks. Banking stock prices are often used by economic analysts as a reflection of the Indonesian economy, this is because banking stock prices are one of the largest sectors in the Indonesian capital market. However, since the discovery of the Covid-19 outbreak in Indonesia in March 2020, banking stock prices have fallen drastically. Since then, the movement of banking stock prices has continued to fluctuate and be uncertainty. This study will forecast banking stock prices using BBCA, BMRI, and BBRI stock price data by adding an intervention variable, namely the time the Covid-19 outbreak was discovered in Indonesia. In this study, we will compare hybrid model of the univariate time series and Artificial Neural Network known as ARIMAX-NN with hybrid model of the multivariate time series and artificial neural network as VARX-NN. The results of this study show that hybrid VARX-NN model produces a smaller RMSE value than ARIMAX-NN model in BBCA, BMRI and BBRI.