An examination of the weak form of the Efficient Market Hypothesis within the context of the NASDAQ Composite Index: A test of the forecasting abilities of artificial neural networks

Since far back in the annals of recorded history, man has feverishly attempted to find ways to predict future events. One area in which this quest has been especially prevalent is in the pursuit of reliable and accurate techniques for predicting the future movement of financial markets. Eugene Fama's Efficient Market Hypothesis is the foundation of most of the stock market research conducted over the past 40 years. During this time there has been considerable investigative testing of the validity of this hypothesis. Fama presents his hypothesis in three forms: the weak form; semi-strong form; and strong form. This dissertation examines the weak form of Fama's hypothesis, which in essence postulates that there is no information relative to the past that can appreciably forecast future security prices.; Traditional finance has attempted to utilize linear regression modeling to create forecasting models. This is a logical step in the right direction. This study examines the application of an artificial neural network model to the prediction of the future price of the NASDAQ Composite Index as compared to a standard multiple linear regression model forecast. The research highlights just how much the behavior of investors, however irrational and non-linear, affects the movement and direction of the NASDAQ Composite Index.; The research in this area to date was predominantly performed during the 1990s. This study extends this research into the new millennium. The new millennium period is deemed significant because of the range of market activities that transpired during this period, including a variety of stock market bubbles, corporate scandals, and economic uncertainties, along with geo-political unrest.