The Analysis Of Medical Time Series And Its Application To Predictions

In the practice of medical research, some measurements are got dynamically in order to find the laws that control the regularity inside. Further more, future values of the concerned variable may be predicted at given confident levels. In the theoretic view, regression models that include the causality between the variable and other influential factors are good at prediction as their foundations are based on more adequate information at least in formality way. However, the influential factors are often too complicated to be measured. Sometimes the patterns of influence are not easy to be inducted exactly. In the situation above, the theory of time series is an efficient tool to process the data. The models will be founded on the series itself. The information of regularity is supposed to be included in the models as much as possible and this insures the reliability of prediction.The modeling and prediction of time series are both beneficial to individuals and groups. To an individual, from early slight pathological changes, later severe changes are supposed to be predicted and the proper treatment may be conducted in time. To a group, pathogenic factors and incidence levels are supposed to be predicted and appropriate interventions are adopted in an efficient way and at proper time.The topics in the thesis are totally produced from medical statistical...