The Application Research Of Fuzzy Discriminatory Analysis In The Forecast Of PIH

Pregnancy-induced hypertension (PIH) is a special disease in gestation. The incidence of this disease is 9.4 percent in our country. And it is from 7 percent to 12 percent. This disease affects the health of mothers and infants very seriously, it is the main reason of the death of lying-in women and infants^[1], and the reason of this disease is unknown. So we can not cure based on reason, the main means in clinic measure is curing early and heteropathy. Thus it can be seen that setting an authentic model to forecast PIH, control state of an illness and reduce harm of mothers and infants. Now the main methods of PIH forecast is single target(single factor) forecast overseas and domestically, its sensitivity and differentia is low or not steady, so the value of forecast still have a insufficiency^[2].the pathogeny of PIH maybe is a integration of many factor^[3]. So forecast by an integration of many factors may advance the value of the forecast method. Now the study of forecast by an integration of many factors is very separate, and relate to very a few factors, and there are not a estimate of forecast effect, I advances fuzzy discriminatory analysis to solve above question, diagnose PIH by forecast.This paper introduced the concept of the PIH, research background, and analyzed the importance of forecast PIH, then introduced the research present condition of PIH forecast, and research content of this paper. The emphases of the paper is advancing fuzzy data character and setting fuzzy discriminatory analysis model, we design and actualize SAS system by Object Oriented method to prove the validity of fuzzy discriminatory analysis model.At last, forecast PIH by discriminatory analysis in multivariate statistical analysis and fuzzy discriminatory analysis based on a mount of real data, analyzed the result of this two methods, prove fuzzy discriminatory analysis is more correct then discriminatory analysis in multivariate statistical analysis, and indicate the research content we can do in the future.