Stock Market Prediction in the Presence of Coronavirus Crash

Since the stock market crash in 1929, the Coronavirus Crash was the fastest and the most devastating crash in the United States financial market. Besides that, thanks to artificial intelligence progress, time series forecasting has been profoundly improved in recent years. This paper analyzes whether predicting the market with high precision is possible with such a worldwide crisis and such progress. The mentioned question is answered by developing a deep learning network to predict the S&P 500 index. The data in this paper consist of the daily opening prices from 4 January 2010 to 16 October 2020. Firstly, the Wavelet analysis is selected to denoise the training data to reduce the complexity of the problem. After noise reduction, the data is used as the input variable of Long-Short-Term-Memory (LSTM), a recurrent neural network (RNN). Finally, Mean Squared Error (MSE) is calculated to evaluate the accuracy of the model, and the results show that despite the sudden crash, the algorithm can predict the market with the MSE of 4.8733e-04.

Keywords:

COVID-19 , Stock Market Prediction , Wavelet , LSTM , Soft Computing