Novel non-destructive machine learning-assisted spectroscopic method for quality control of bleached kraft pulps

Data related to a submitted manuscript 'Novel non-destructive machine learning-assisted spectroscopic method for quality control of bleached kraft pulps'. The datasets contain raw NIR and Raman spectra, extractive contents in birch and conifer pulps and pre-processed NIR and Raman spectra (Savitzky-Golay smoothing and standard normal variate) and the datasets after outlier removal following Hotelling's T2 and F-residuals scores (95% confidence interval). The smoothing and normalization were performed in Orange Data Mining, while outlier removal in Unscrambler. The NIR and RAman spectral data was used for building PLS models for prediction of various extractive contents in birch and conifer pulps.