Modeliranje pojavnosti endemske nefropatije logističkom regresijom

U ovome radu modeliramo pojavnost endemske nefropatije logističkom regresijom. Promatramo modele nastale "FORWARD", "BACKWARD" i "STEPWISE" metodom selekcije, te zadani model i njegove varijacije. Statistička obrada podataka rađena je u statističkom paketu SAS Studio. U svakom modelu dobivenom SAS-ovim metodama selekcije kao rizični faktori za razvijanje endemske nefropatije pojavljuju se duljina života u endemskom selu, razina \(\alpha\)1 mikroglobulina te pozitivna obiteljska anamneza na endemsku nefropatiju.In this thesis, we are modelling prevalence of endemic nephropathy with logistic regression. We are observing models which are created with "FORWARD", "BACKWARD" and "STEPWISE" selection methods, and also demanded model and its variations. Statistical analysis was done in statistical software SAS Studio. In each model created with SAS selection methods, living duration in endemic village, level of $\alpha$1 microglobulin and positive family history on endemic nephropathy are appearing as risk factors for developing an endemic nephropathy