Abstract
Data presented in this article are related to the research article entitled "Near Infra-red spectroscopy quantitative modelling of bivalve protein, lipid and glycogen composition using single-species versus multi-species calibration and validation sets" [1]. Band width selections were determined using a data-driven approach to modelling Near Infra-red Spectra (NIRS) of protein, lipid and glycogen content in bivalves. Models were produced for single species and combined species of Saccostrea glomerata, Ostrea angasi, Crassostrea gigas, Mytilus galloprovincialis and Anadara trapezia. Band width selection was undertaken using Fourier wavelet transformation coupled with a genetic algorithm (GA) to aggregate adjacent wavelet bands to select the minimum number of IR bands that were consistently identified in the majority of individual spectra.