CAE Proceedings on International Conference on Communication Technology |
Foundation of Computer Science USA |
ICCT2015 - Number 6 |
September 2015 |
Authors: Shashank Iyer, Nisarg R. Kamdar, Bahar Soparkar |
7f57f905-fbb3-4ef0-b260-bbf72d5efbcc |
Shashank Iyer, Nisarg R. Kamdar, Bahar Soparkar . Stochastic Signal Modeling Techniques for Stock Market Prediction. CAE Proceedings on International Conference on Communication Technology. ICCT2015, 6 (September 2015), 40-45.
Autocorrelation measures the degree to which a current variable is correlated to the past values. Autocorrelation can be measured by running a regression equation. The study employs this temporal correlation that exists between the various stock markets related variables to predict future trends and prices, using two stochastic signal modeling processes. Data for stocks listed on the NASDAQ was scraped from the Yahoo! Finance website. Autoregressive (AR) and Autoregressive Moving Average (ARMA) techniques have been used to predict the next day's closing price using a time series input of the previous L days. Autoregression models the dependence of the variable to be predicted with its own lagged terms while Autoregressive Moving Average builds on Autoregression by allowing for the introduction of the Moving Average model which includes lagged terms on the residuals. The mean square error of the two was compared. The study concludes that the two models should be used in consonance for accurately modeling the magnitude and the direction of the movement in the variable to be predicted.