This work investigates the potential of Threshold Models in interpreting some characteristics of Financial Markets through an application to three stocks.we will use the UP/DOWN and trading volume of stocks as Threshold Variable to build a threshold autoregressive model.In the stock market, the price of the stock is influenced by various factors.Trading volume and the UP/DOWN of stocks are tow very important factors. Generaly speaking , business both sides obtain the confirmation to the price through the size of the trading volume .The confirmation is large , the trading volume is large; The confirmation is little , the trading volume is small.This kind of market behavior of both sides demonstrates one such trend law :when the price increases, the quantity increases;when the price falls, the quantity reduces.In finance,the majority of data are non-linear, this obtained people's recognition. The question which must be solved is that we select which type of non-linear model to fit the dates. The performances of the TAR model are evaluated in comparison with other competitive models,such as a linear AR,a GARCH,a CHARMA model. The TAR model is not only capable of capturing the various characteristics of volatility ,but is in particular the only one which can distinguish a persistent shock of the market from an extraordinary shock.Its superior performance is even more evident in particularly turbulent moments of the market.From a modeling point of view the innovative procedure introduced by Tasy is employed here.He pointed out this methodology in order to make Threshold Models tractable in application,since previous techniques were expensive. |