A new stock price prediction method using covariance information
Date
2017
Authors
Ghorbani, Mahsa, author
Chong, Edwin, author
Journal Title
Journal ISSN
Volume Title
Abstract
Stock price prediction is one of the most challenging problems in finance and is receiving considerable attention from researchers. The literature provides strong evidence that prices can be predicted from past price data as well as other fundamental and macroeconomic variables. We propose a filtering operation using covariance information in order to predict future stock prices. We use daily historical price data for Generals Electric Company to illustrate our method, which shows promising results in terms of the estimation performance and volatility.