| dc.description.abstract | The objective of this study was to apply NIR method to analyze modified cassava flour (MOCAF) accurately, more simple, requiring no chemicals, and quickly predict the chemical compositions (moisture, ash, pH, amylose contents). The wavelength ranges to predicting MOCAF chemical compositions, from 1000 to 2500 nm (4000-10000 cm-1 at intervals of 4 cm-1). Prediction results represent the reflectance data (R) of 70 samples with 1500 wavelength data, whereas the absorbance data (A) was obtained with transformation log (1/R). Calibration method which used in this study are principal component regression (PCR) and partial least squares (PLS). Data treatment on the reflectance and absorbance spectrum curve estimation MOCAF chemical compositions, among others: smooth average 3 points, second derivative Savitzky-Golay 9 points, and combination both of them. A number of 70 MOCAF were used as samples. Samples were divided into two phases: ± 45 samples (2/3 of total samples) for developing calibration equation and ± 25 samples (1/3 of total samples) for performing validation. NIR data analysis result shows that PLS method with NIR reflectance data and the smooth average 3 points is the best method of calibration and data treatment to predicting moisture contents of MOCAF. Prediction of ash, pH, and amylose contents of MOCAF best obtained with the PLS method, the NIR absorbance data, and combination data treatment of smooth average 3 points and second derivative Savitzky-Golay 9 points. The standard error of prediction (SEP) and coefficient of variability (CV) respectively were 0.49% and 4.2% for moisture contents; 0.02% and 2.8% for ash contents; 0.02 and 0.4% for pH contents; and 0.39% and 1.3% for amylose contents. | en |
