期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2015
卷号：3
期号：9
DOI：10.15680/IJIRCCE.2015. 0309079
出版社：S&S Publications
摘要：Volatility in stock market refers back to the movement of stocks; generally it may be defined as thedanger in investing on shares. Volatility is measured as standard deviation and variance of Closing prices. Forecastingvolatility has been a first-rate challenge in economic market and plenty of researchers are working on it. The main goalof this paper is to forecast volatility with a high accuracy. The volatility is calculated making use of standard deviationsof returns from the closing prices. This research focus is to forecast volatility with high accuracy by using 10 differenttechniques that involves both machine learning techniques and time series techniques. The selected techniques foranalysis are Naïve Forecast and Neural Networks, ARIMA, ARFIMA, BATS Forecast, TBATS Forecast, BoxCoxForecast, Rand walk Forecast, normal method and Holt Winters forecasting techniques. The results of all thetechniques are in comparison with other techniques to find an accurate forecasting technique. The excellent forecastingmethod is shortlisted by evaluating the error outcome of all of the forecasting procedures with error measuringparameters such as ME, RMSE, MAE, MPE, MAPE and MASE. ARIMA technique is more accurate volatilityforecasts for subsequent 10 days.
关键词：Forecast; Volatility; Machine Learning; Times series
