Seleksi Panjang Gelombang yang Efisien pada Nir Spectroscopy untuk Pengukuran Kandungan Padatan Terlarut Buah Salak Pondoh Menggunakan Model Forward Interval PLS (FiPLS)

In this research, the feasibility of using NIR spectroscopy coupled with forward interval partial least squares (FiPLS) regression for nondestructive soluble solids content (SSC) determination in Pondoh snake fruit was investigated. Spectra of snake fruit were acquired using a VIS-NIR USB4000 portable spectrometer in absorbance mode. The SSC was measured using a digital refractometer. The calibration and validation models were developed using the FiPLS regression using original spectra. The result showed that the optimal combinations of 7 spectral intervals among 11 intervals which selected by FiPLS yielded good results (RMSEC = 0.854783%, rc= 0.96, RMSECV = 0.899828%, rv = 0.95).  Keywords: soluble solids content, absorbance mode, nondestructive method, snake fruit, FiPLS regression, NIR spectroscop