Robust Regression Models for Identifying Outliers in Medical Data
Abstract
Before applying any multivariate statistical analysis, it is important to determine whether outliers are present in the dataset. In regression analysis, the presence of outliers in the dataset can strongly mislead the classical least squares estimator and lead to unreliable results. In this paper, we prove the minimum covariance determinant estimator, which is commonly applied in a robust statistic to estimate location parameters and multivariate scales. These ideas can be used to robust distance of Sign method, Mahalanobis distances and Cook’s distances to identify outliers. The intent of this robust regression study is to provide a behavior of outliers in linear regression and to compare robust regression methods.
Keywords: Robust distance, sign method, Cook’s distance, outliers
Cite this Article
Sampath G, Senthamarai Kannan K, Manoj K. Robust Regression Models for Identifying Outliers in Medical Data. Research & Reviews: Journal of Statistics. 2018; 7(1): 69s–76sp
Full Text:
PDFDOI: https://doi.org/10.37591/rrjost.v7i1.831
Refbacks
- There are currently no refbacks.