Near-infrared spectral data and physicochemical attributes for chicken meat quality classification

This dataset contains near-infrared (NIR) spectral measurements and physicochemical attributes of chicken breast meat samples used for poultry meat quality classification. The dataset is associated with the article Machine Learning Applied to Near-Infrared Spectra for Chicken Meat Classification, published in Journal of Spectroscopy in 2018.

The spreadsheet includes spectral values collected from minced chicken breast fillets in the 400-2498 nm range at 2 nm intervals, totaling 1050 wavelengths. Additional measured attributes include pH, CIELAB color coordinates, water holding capacity (CRA/WHC), texture, chroma, hue, and quality class labels. Class labels are encoded as: 1 = PSE (pale, soft, and exudative), 2 = normal, 3 = pale, and 4 = DFD (dark, firm, and dry).

The local file inspection identified 158 samples with the following class distribution: 24 PSE, 41 normal, 86 pale, and 7 DFD. These values should be validated by the data owners before publication.