Multiple regression for single index model

Kenichi Satoh, Megu Ohtaki
2002 Proceedings of the Symposium of Japanese Society of Computational Statistics  
A simple method based on SIR (Sliced Inverse Regression) is proposed to exp}ore an EDR vector fbr the single index model. We avoid the principle component analysis step of the original SIR by using two sample mean vectors in two slices of the response variable and their difference vector. The methodology can be effeetively applied even when the number of coyariates is Iarge since it requires no matrix operation or iterat]ive calculation.
doi:10.20551/jscssymo.16.0_51 fatcat:dqxiqo7csvf4feoxg4c6br5cbu