a new approach to the predictability analysis of tehran stock market

using nonlinear mathematical analysis, on the data obtained for.time seriesof shahd-iran price during 3.5 years, the characteristics,of the process associated with this, is analysed. analysing the behaviour of the time series associated with returns is indicative ofits short-term predictability nature. however, employing analysis regarding the correlation dimension estimate. it is indicated that, only time series ofprice (returns) is not adequate for prediction and other appropriate variables must also be used. the correlation dimensionestimate indicates the complexity ofthe prediction model also, largest lyapunov exponent analysis, reveals a weakly chaotic behaviour and indicates that price data cannot be used in the prediction process after a certain time.