Articles | Volume 11, issue 2
Research article
02 Mar 2018
Research article |  | 02 Mar 2018

Evaluation of linear regression techniques for atmospheric applications: the importance of appropriate weighting

Cheng Wu and Jian Zhen Yu

Model code and software

Scatter Plot C. Wu

Aethalometer data processor C. Wu

Histbox C. Wu

Short summary
A new data generation scheme that employs the Mersenne twister (MT) pseudorandom number generator is proposed to conduct benchmark tests on a variety of linear regression techniques. With an appropriate weighting, Deming regression (DR), weighted ODR (WODR), and York regression (YR) are recommended for atmospheric studies when both x and y data have measurement errors. An Igor-based program (Scatter Plot) is developed to facilitate the regression implementation.