This research is concerned with the analysis of a set of 1000 primary data obtained from Cell Phone user on switching behavior.The main objectives of our study were to fit the markovian model and Logistic regression model.Our response variable was switching behavior and we have 16 independent variables. The test result shows that the markovian and logistic regression models were well fitted to the data. With the help of transition probability matrix we went through the predictions of what would be the cell phone connection among all operators after one, two, or more times switching.By the estimated limiting behavior of the transition probabilities of the four cell phone companies between the times we interpreted that at the coming fifth times for past and present connections and coming seventh times for present and future connections, the cell phone subscriber s distribution will be stationary i.e., no changes will take place afterward. This phenomenon was also examined with the help of the test for the stationarity.And at 5% level of significance the logistic regression model is well fitted with the ten significant covariates.

Autorentext

MS in Applied Statistics from University of Dhaka,Bangladesh. Black Belt in SIX SIGMA Business Process Improvement Methodology. Head of Process Performance Management. Business Process Management Department. GrameenPhone Limited,Bangladesh.