Simulation and prediction of clinical states of leukemia by the auto-regression model

The Auto-Regression method was applied to the modelling of leukemic states. The number of two kinds of normal cells and of a kind of abnormal cells in the peripheral blood were taken as output variable. The model obtained explained the trend of the normal cells in the prediction, while it could not satisfactorily follow the time course on the knowledge on the drug effect to the abnormal cells was constructed. The conbination of these two models seems to be feasible for describing the dynamics of leukemic states.