Influence of data pre-processing techniques for PLSR model to predict blood glucose by NIR spectroscopy

NIR diffuse reflectance spectroscopic spectra can be mathematically modelled to extract quantitative information by suitable multivariate calibration models. The analysis of spectral data becomes complex as the is more prone noise due light scattering and baseline effects. These errors reduces robustness reliability developed model. Hence pre-processing most important aspect in analysis. Different mathematical transformations are applied remove present data. This work focuses on various empirical techniques like correction, multiplicative scatter correction (MSC), robust MSC, extended signal (EMSC), orthogonal (OSC) (-log R) followed standard normal variate (SNV) for Partial Least Square Regression (PLSR) model prediction blood glucose non-invasively. performance PLSR acquired (raw) same subjected different analyzed. complexity evaluated terms number latent variables (LVs) required build obtained mean square error after cross validation. study utilizes collected from 207 subjects a diabetic center using Diffuse Reflectance Spectrometer (DRS). analyzed results show that based SNV found perform well with reduced minimum estimated 0.23 mg/dl among other techniques. Keywords: (OSC), (SNV),