Extraction of most effective wavelength of mulberry juice (Morus Alba Var. nigra L.) spectra using different PLSR models

In recent years, application of UV/NIR spectroscopy as a non-destructive technique combined with chemometric methods has been widely noticed for quality assessment of food and agriculture produces. Mulberry is a native fruit of Iran and there are limited information about its quality indexes. In this paper, backward interval partial least-squares (BiPLS) and interval partial least-squares (iPLS) were utilized as wavelength selection methods in UV/NIR region to determin fruit juice internal qualities (soluble solid content (TSS), Titratable Acidity, Anthocyanine Content and Ascorbic Acid). In order to find the appriptare number of components for each juice parameter Cross-Validation method was used.BiPLS method with 15 factors of 20, compares to PLS and iPlS, had the best performance in prediction of fruit juice parameters SSC (RMSEC = 0.076, rc = 0.96 and RMSEV = 0.03, rv = 0.96), TA (RMSEC = 0.014, rc = 0.96 and RMSEV = 0.013, rv = 0.98), AA (RMSEC = 0.053, rc = 0.87 and RMSEV = 0.0008, rv = 0.93) and AC (RMSEC = 0.0007, rc = 0.97 and RMSEV = 0.0006, rv = 0.95).