Multiplicative Scatter Correction Improves Near-Infrared Spectroscopy–Based PLS Models for Acetic Acid Quantification in Jamblang Vinegar

Jamblang (Syzygium cumini L.) vinegar is a potential functional fermented product that requires rapid and reliable quality assessment. This study developed a near-infrared spectroscopy (NIRS)-based method combined with partial least squares regression (PLSR) to quantify acetic acid content in jamblang vinegar. Vinegar samples representing a wide range of acetic acid concentrations were obtained through varied fermentation conditions and analyzed using a reference enzymatic method. NIR spectra were preprocessed using Multiplicative Scatter Correction (MSC) to reduce scattering effects caused by sample heterogeneity. PLSR models constructed from MSC-treated spectra showed improved predictive performance compared with uncorrected spectra, achieving a coefficient of determination (R2) of 0.995 and a residual predictive deviation greater than 2.3. Regression coefficient analysis identified wavelength regions associated with O–H and C–H overtone vibrations as key contributors to acetic acid prediction. The results demonstrate that NIRS coupled with MSC-enhanced PLSR provides a rapid, non-destructive, and accurate approach for acetic acid determination in jamblang vinegar, supporting its application for routine quality control and fermentation monitoring.