Artificial Neural Network-Based Machine Learning Approach to Stock Market Prediction Model on the Indonesia Stock Exchange During the COVID-19.

The global COVID-19 pandemic has caused panic. In addition, it disrupted life and economic activities around the world. Prediction of the stock market during the COVID-19 pandemic became a major challenge because the data was not stationary, random, and complex nonlinear system. For this reason, an in-depth study of the following trends is required to develop an adequate predictive model to predict the stock market during the pandemic. This study designs a stock market prediction model during the COVID-19 pandemic on the Indonesia Stock Exchange using a deep learning approach based on artificial neural networks. The object of this research is the pharmaceutical industry in the health sector listed on the IDX. The input variables are the proposed model for predicting stock prices with daily stock price movements, including COVID-19 trend indicators, and the government's response tightness index to COVID-19 in Indonesia. The study results show that all proposed model systems achieve highly accurate forecasting for the stock market price prediction with MAPE 10%. Model 6-20-20-1 is the best model of all tested models, with MSE = 0.00055, RMSE = 0.007418, and MAPE = 1.17%. [ABSTRACT FROM AUTHOR]