NIRS Calibration for retting ability of soaked cassava roots at NRCRI, Umudike, Nigeria

The retting roots of soaked cassava roots were evaluated for potential good calibration and prediction equation models using handheld NIRS equipment. Seventy (70) cassava genotypes from Umudike and Otobi locations from Crossing block and Advanced Yield Trials (AYT) of NextGen were assessed for retting ability. Near-infrared spectra models were developed in “Win ISI 4 Project Manager”, by using the modified partial least squares (MPLS) regression and cross-validation techniques. Calibrations were done and the coefficient of determination in calibration for retting ability (R²cal) was 0.84 with SECV of 2.4 and SEC of 2.19. Also, the performances of the prediction models were tested using an independent set of 20 independent retting cassava root samples. A good prediction coefficient (R2 pred) and low standard error of prediction (SEP) was obtained 0.86 and 1.16 respectively.