Premium
Empirical performance of nonparametric regression over LRM and IGRM addressing influential observations
Author(s) -
Khan Javaria Ahmad,
Akbar Atif
Publication year - 2019
Publication title -
journal of chemometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.47
H-Index - 92
eISSN - 1099-128X
pISSN - 0886-9383
DOI - 10.1002/cem.3143
Subject(s) - nonparametric regression , nonparametric statistics , bayesian multivariate linear regression , outlier , kernel regression , multivariate statistics , statistics , econometrics , regression analysis , kernel (algebra) , mathematics , computer science , linear regression , kernel method , linear model , artificial intelligence , support vector machine , combinatorics
Nonparametric regression is commonly used for summarizing the relationship between variables without requiring the assumptions of model. Generalized linear model and linear regression model are usually used to examine the relationship of variables, but both are badly affected by influential observations. Due to this, detection and removal of outliers attain a lot of attention of researchers to obtain reliable estimates. We focus on such robust technique whose performance is acceptable in the presence of outliers. The present article empirically compared the performance of linear regression model and generalized linear model with multivariate nonparametric kernel regression. Here, multivariate nonparametric kernel regression is used with Gaussian kernel and six different bandwidths on Aerial biomass data. The performance of nonparametric regression with Bayesian bandwidth was found to be better as compared with other methods.