MLNIRdata: NIR spectra and densities of hydrocarbon mixtures (208 samples) for property prediction.

This dataset contains 208 near-infrared spectra of hydrocarbon mixtures with measured densities. Data are provided as CSV matrices (raw and Savitzky–Golay derivatives), axis files, and a MATLAB MAT bundle for convenient loading in Matlab, Python, Julia, or R. It supports chemometric and machine-learning tasks such as regression and benchmarking of property prediction models.