Classification of Mango by Near Infrared Diffuse Reflectance: Comparison of Methods Used in Predicting the Sugar and Acid of Mango

Abstract The classification of mango ( Mangifera indica var. gedong) into their taste, sweet, sweet-sour, and sour which were traditionally judged by aroma and experiences is becoming increasingly more difficult due to the cross breeding and to the International's consumer demand. The NIR system was developed and applied to 200 mango samples at the wavelength range from 1400 - 2000 nm. The samples were separated into two main group i.e. calibration data and validation data. The stepwise method was used to select the best wavelength for calibration equation. This calibration equation was then validated using the second data and also compared with the Principal Componen Analysis (PCA)of the selected samples. It was found that the standard error of prediction (SEP) of the validation data either using the multiple regression or PCA shows an agreeable correlation.