Forecasting Stock Indexes by Reinitiated Singular Spectrum Analysis Approach

Authors

  • Kong Hoong Lem, Woan Lin Beh

Abstract

This paper proposed a reinitiated singular spectrum analysis (rSSA) approach for stock index forecasting.  The approach was experimented on some international stock indexes. Forecast performance was benchmarked against the conventional Autoregressive Integrated Moving Average (ARIMA) models using the agreement index?. In addition, other common performance metrics such as the weighted mean absolute percentage error (wMAPE) and the root mean of squared errors (RMSE) also served as reference.  For the data tested, the rSSA forecast outruns those from ARIMA family. Reinitiated SSA is a potential alternative for stock market index forecasting.

Downloads

Published

2020-05-12

Issue

Section

Articles