A suitable statistical model has been explored for the investors as well as the researchers to resolve the future estimation of share volume by using daily stock volume data from Dhaka Stock Exchange (DSE). The dail...A suitable statistical model has been explored for the investors as well as the researchers to resolve the future estimation of share volume by using daily stock volume data from Dhaka Stock Exchange (DSE). The daily volume data from the June 1, 2004 to April 19, 2010 were retrieved from DSE website as a secondary data source. The Maximum Likelihood---Autoregressive Conditional Heteroskedasticity (ARCH) (Marquardt) method has been applied to construct the models for the stock volume data of DSE by using statistical package software E-Views of verson-5. First of all, an "Auto Regressive Integrated Moving Average (ARIMA) model" was fitted and observed that heteroscedastic volatilities were still present there. To eliminate this dilemma, ARCH class of volatility models has been used and finally the ARIMA with EGARCH model has been explored. Findings of this study have recognized that ARIMA with EGARCH model implies low mean square error, low mean absolute error, low bias proportion, and low variance proportion for share volume data with comparing to other models. Hence, the modelling concept established in this study would be a decisive study for the investors as well as the researchers.展开更多
文摘A suitable statistical model has been explored for the investors as well as the researchers to resolve the future estimation of share volume by using daily stock volume data from Dhaka Stock Exchange (DSE). The daily volume data from the June 1, 2004 to April 19, 2010 were retrieved from DSE website as a secondary data source. The Maximum Likelihood---Autoregressive Conditional Heteroskedasticity (ARCH) (Marquardt) method has been applied to construct the models for the stock volume data of DSE by using statistical package software E-Views of verson-5. First of all, an "Auto Regressive Integrated Moving Average (ARIMA) model" was fitted and observed that heteroscedastic volatilities were still present there. To eliminate this dilemma, ARCH class of volatility models has been used and finally the ARIMA with EGARCH model has been explored. Findings of this study have recognized that ARIMA with EGARCH model implies low mean square error, low mean absolute error, low bias proportion, and low variance proportion for share volume data with comparing to other models. Hence, the modelling concept established in this study would be a decisive study for the investors as well as the researchers.