作者: Beata Walczak , Wolfhard Wegscheider
DOI: 10.1016/0003-2670(93)85264-K
关键词:
摘要: Abstract The combinations of classical bilinear models and neural nets, extended to net on residuals from partial least squares (PLS) are discussed. performances principal component regression (PCR), PLS, networks (NN), analysis (PCA)-NN PLS residuals-NN compared simulated data, near-infrared data quantitative structure-activity relationship data.