Modeling stock prices
by MSc talk by Xuan Li
Abstract: Recently, most people are interested in movements of the stock and index prices. As a result, many researchers are tried to use different techniques to predict the future movements of both stock prices and index prices. This is because they believe that those prices may give some directions of how stocks perform from the data in the past.
Generally, this thesis will demonstrate several different models to predict the future stock prices. Those model interpretations include smoothing techniques such as exponential smoothing and Holt methods, and various types of Generalized Autoregressive Conditional Heteroskedasticity models, a.k.a GARCH models. Furthermore, this thesis will look at the model of multivariate distributions to inference multivariate exponential smoothing and GARCH models via Copula technique. Finally, after performing the models from the historical data of some stocks, we will observe some results of whether those model predictions are good or not via graphical representations and statistical techniques.