1：Experimental Design and Method Selection:

The study used a data fusion approach combining NIR and MIR spectroscopy with chemometric algorithms (Si-PLS and GA-PLS) for quantitative analysis of polyphenols in Chinese dates. The rationale was to improve prediction accuracy by integrating information from both spectroscopic techniques.

2：Sample Selection and Data Sources:

80 Chinese date samples were purchased from local suppliers in Henan, Xinjiang, Shanxi, Shandong, and Hebei provinces. Samples were prepared by puff drying, grinding, and sieving to 500 μm mesh, then dried at 40°C for 24 h to eliminate moisture interference.

3：List of Experimental Equipment and Materials:

Equipment included an Antaris? II NIR spectrophotometer (Thermo Electron Company), an FT-MIR spectrophotometer (Nicolet iS50, Thermo Scientific), a puff dryer, an electric grinder (QE-100, Zhejiang YiLi Tool Co., Ltd), and standard laboratory materials for polyphenol extraction and analysis (e.g., Folin-Ciocalteu reagent, Whatman No. 1 filter paper).

4：Experimental Procedures and Operational Workflow:

NIR spectra were acquired in reflectance mode with an integrating sphere, averaging 16 scans per spectrum over 4000-10000 cm?1. MIR spectra were acquired using ATR accessory with 32 scans over 4000-650 cm?1. Total polyphenol content was determined colorimetrically using Folin-Ciocalteu reagent. Spectra were preprocessed with SNV and de-trending, then divided into calibration (54 samples) and prediction (26 samples) sets. Si-PLS was used for spectral interval selection, and GA-PLS for variable optimization.

5：Data Analysis Methods:

Data analysis was performed in Matlab (Version 7) using PLS, Si-PLS, and GA algorithms. Performance was evaluated using R2, r2, RMSECV, RMSEP, bias, and RPD values.