Forecasting the adjusted Dow Jones Industrial Average index
No Thumbnail Available
School of Business | Bachelor's thesis
(Mikkeli) Bachelor’s Program in International Business
37 + 2
AbstractObjectives The main objectives of this study were to examine the performance of a Nonlinear Autoregressive Exogenous Artificial Neural Network in forecasting the daily adjusted closing prices of the adjusted Dow Jones Industrial Average index. The performance of the Neural Network is compared with the traditional Capital Asset Pricing Model. This allows the evaluation of both model’s capabilities in forecasting as well as testing the underlying assumptions in Capital Asset Pricing Model. Summary The literature review views the use of artificial intelligence in finance, concentrating on their use in financial forecasting while also discussing the Capital Asset Pricing Model and its potential for forecasting. Subsequently a Neural Network is constructed, trained and applied to the task of forecasting, while simultaneously a model for forecasting the index’s closing prices with the Capital Asset Pricing Model is developed. Finally the results from the two models are compared and the implications of the results discussed. Conclusions The findings show that both the Neural Network developed and the Capital Asset Pricing Model can be used to forecast the adjusted stock index up to half a year forward. Although both models can be used the Neural Network outperforms the Capital Asset Pricing Model significantly, as measured by their respective mean squared errors of forecasts. Regardless, the assumed relationship between risk and return in the Capital Asset Pricing Model does not appear to hold true.
Thesis advisorStepanov, Roman
artificial intelligence, machine learning, financial forecasting, neural networks, finance, forecasting, forecasting techniques