Abstract: A process for estimating a critical dimension of a trench formed by etching a substrate. First, a regression model is constructed for estimating the critical dimension, in which principal component loadings and principal component scores are also calculated. Next, a substrate is etched and spectral data of the etching are collected. A new principal component score is then calculated using the spectral data and the principal component loadings. Finally, the critical dimension of the trench is estimated by applying the new principal component score to the regression model.