The ability of spectrum autocorrelation models to predict the lycopene concentration in foods through visible spectroscopic data

We developed a novel computerized approach based on lag-k autocorrelation coefficients (LCCs) and linear models (LMs) to estimate the concentration of lycopene in foods by the spectroscopy. The LCCs were calculated using the data obtained using whole visible scans from 400 to 600 nm (vide supra) of lycopene standards and food samples (ketchup, tomato juice and tomato sauce). The chaotic parameter (CP) was then transferred into a LM to estimate the concentration of lycopene compound. The integrated LCC/visible spectroscopy method developed can be considered as a satisfactory analytical technique able to estimate lycopene concentration in food samples in a fast accurate way, with a mean prediction error lower than 5.7% and a mean correlation coefficient higher than 0.957.