On the Predictability of Price Fluctuations in Tehran Stock Exchange A Correlation Dimension Estimation Approach

This paper employs a general non-linear analysis tool to analyse the nature of time series associated with the price (returns) of a particular company in Tehran Stock Exchange. It is shown that the behavior of the process associated with the price (returns) time-series of this company is weakly chaotic, and due to the non-random behavior of the process, short term prediction of stock price is possible. It is also shown, using the correlation dimension estimation analysis, that a modeling of the price fluctuations based solely on the price data is insufficient to establish a model for future price prediction and that other variables involved in the process must be accounted for.