Models for predicting juvenile hypertension

Abstract

Aim of present study was to create some models to predict juvenile hypertension. Data was taken from longitudinal study of Juvenile hypertension in Lithuania. The first survey of a random sample of Kaunas schoolchildren born in 1964 was carried out in 1977. This population was re-examined in 1989, in 1997 and the last survey took place in 2002. There were created models for predicting juvenile hypertension using real data. The models were created using logistic regression, nonparametric discriminant methods (kernel and k-nearest neighborhood) and system SAS. The best results were received using nonparametric discriminant analysis.

Keywords

• logistic regression
• nonparametric dicriminant analysis
• juvenile hypertension
• prediction