For the models (LDA and QDA) using all five predictors, check the assumption of multivariate normality for the three sets of predictor variables (split by origin).

MX_American = NewAuto %>% filter(origin == 1) %>% dplyr::select(mpg,cylinders,displacement,horsepower,weight)MX_European = NewAuto %>% filter(origin == 2) %>% dplyr::select(mpg,cylinders,displacement,horsepower,weight)MX_Japanese = NewAuto %>% filter(origin == 3) %>% dplyr::select(mpg,cylinders,displacement,horsepower,weight)mhz(MX_American)$mv.testmhz(MX_European)$mv.testmhz(MX_Japanese)$mv.test

> mhz (MX_American )$mv . test Statistic p-value Result 3. 657 NO > mhz (MX_European) $mv . test Statistic p-value Result 2. 8197 NO > mhz(MX_Japanese)$mv . test Statistic p-value Result 3. 0102 S NO