I was interested in both cellular automata(NKS) and chaos theory capability of forecasting for stock exchange. after some research in 2005 I find out NKS is out of my league and I chose chaos analysis. Here is abstract of my thesis.
Accuracy and efficiency of economic forecasting models are strategic and crucial of
business world. Many economists believe that linear models are not efficient enough.
So many researches focus on understanding different economical time series structure
and dynamical models that can fit them.
I try to define the chaos theory, short review on business data investigation and finally
I will investigate the Tehran stock exchange Index for chaotic behavior. I will use
Correlation Dimension, Hurst, Largest Lyapunov Exponent and BDS tests for my
investigation. If the hypotheses do not reject it, it shows that it is possible to develop
a dynamical model for short term forecasting of the market. Even information
memory calculated by Hurst and LLE tests can clear the forecasting limitation. Tests
results show enough evidence to accept the Hypotheses of chaotic behavior but due to
weakness of chaotic tests for economical data, we still need to wait for new tests for a
Keywords: Chaos Theory, Lyapunov Exponent, Hurst, Correlation Dimension,
Economical time series, Stock markets