Скачать или смотреть Quantile

Normally the relationship between two variables x and y is studied using the linear regression equation. Linear regression equation requires normality and homoscedasticity (equal variance) assumptions. When the normality and homoscedasticity assumptions are violated, linear regression estimates are not valid. Quantile regression method overcomes the drawbacks of linear regression and can be applied when the data is skewed and equal variance assumptions are violated. This paper provides an overview of application of quantile regression in clinical research using R and SAS statistical package.