Determination of hydroxyl value of palm-based polyols by partial least squares algorithm using nir spectroscopy

Near infra-red (nir) spectroscopy with partial least square (pls) regression of the results was developed as a rapid method for the determination of the hydroxyl value (hv) for palm-based polyols. the best nir pls calibration models to predict hv were developed in two types of spectral regions,model 1 at 6500-8100 cm-1 and model 2 at 4500-10 005 cm-1. for model 1,the nir-generated calibration plots had coefficients of determination (r2) of 0.991 and a standard error of calibration (sec) of 3.23,while the validation plots had r2 of 0.994 and a standard error of prediction (sep) of 2.96. for model 2,the calibration plot had r2 of 0.997 and sec of 2.32,while the validation plots had r2 of 0.998 with sep of 2.52.