Solving the Spectroscopy Interference Effects of β-Carotene and Lycopene by Neural Networks

Abstract

In this study a new computerized approach and linear models (LMs) to solve the UV/vis spectroscopy interference effects of beta-carotene with lycopene analysis by neural networks (NNs) are considered. The data collected (absorbance values) obtained by UV/vis spectrophotometry were transferred into an NN-trained computer for modeling and prediction of output. Such an integrated NN/UV/vis spectroscopy approach is capable of estimating beta-carotene and lycopene concentrations with a mean prediction error 50 times lower than that calculated by the LM/UV/vis spectroscopy approach (without any previous physicochemical knowledge of the process to be modeled).

Qualitative and quantitative analysis of milk for the detection of adulteration by Laser Induced Breakdown Spectroscopy (LIBS)

“…Although there is a need in the food industry to improve quality and process controls, the application of LIBS technique in food control analysis has not been sufficiently explored and only few studies can be found in literature (Bilge, Boyacı, Eseller, Tamer, & Çakır, 2015;Bilge, Sezer, Eseller, Berberoglu, Topcu, & Boyaci, 2016;Bilge, Velioglu, Sezer, Eseller, & Boyaci, 2016;Mbesse Kongbonga, Ghalila, Onana, & Ben Lakhdar, 2014). The combination of LIBS with chemometric methods offers the possibility to be used in a fast, automatic and on-line manner as has already been demonstrated with successful results for sample classification and quantification (Caceres, Moncayo, Rosales, de Villena, Alvira, & Bilmes, 2013;Curteanu & Cartwright, 2011;Huang, Kangas, & Rasco, 2007;Moncayo, Manzoor, Navarro-Villoslada, & Caceres, 2015;Moncayo, Rosales, Izquierdo-Hornillos, Anzano, & Caceres, 2016;Torrecilla, Cámara, Fernández-Ruiz, Piera, & Caceres, 2008). Some authors have shown that the molecular LIBS signal is dependent on the molecular nature of the sample (Anzano, Casanova, Bermúdez, & Lasheras, 2006;Gregoire, Boudinet, Pelascini, Surma, Detalle, & Holl, 2011).…”

External calibration models for the measurement of tomato carotenoids by infrared spectroscopy

“…The increasing interest in carotenoids as bioactive food components has promoted the development of numerous rapid methods for their measurement in tomato and tomato products, which are the main source of carotenoids in the western diet (Giovannucci, 2002). Techniques such as UV-vis (Choudhary et al, 2009;Pflanz and Zude, 2008;Torrecilla et al, 2008), near-infrared (Clé ment et al, 2008;Kusumiyati et al, 2008;Pedro and Ferreira, 2005), mid-infrared (De Nardo et al, 2009;Halim et al, 2006;Rubio-Diaz et al, 2010) and Raman (Bhosale et al, 2004;Schulz et al, 2005) spectroscopy have been frequently proposed as alternatives to HPLC for the simple, high-throughput and costeffective analysis of tomato carotenoids. The selection of the most adequate spectroscopic method will depend on the objective of the analysis, characteristics of the sample, expected accuracy and specificity, and equipment amenability.…”

“…However, neural networks (Torrecilla et al, 2008) and iterative multiple linear regression analysis (Pflanz and Zude, 2008) have been shown to improve the performance of UV-vis spectroscopy for the measurement of lycopene and b-carotene in regular tomatoes. The analysis of new tomato varieties with complex profiles of carotenoids in various isomeric configurations requires methods with higher degrees of specificity.…”

New Developments in Lycopene Analysis by Spectroscopic and Chromatographic Techniques, Accompanied With Mathematical Modelling

“…New computerized approaches and linear models (LMs) to solve the UV/ visible spectroscopy interference effects of b-carotene with lycopene analysis by neural networks (NNs) have been reported in recent years [236,237]. The data (absorbance values) obtained by UV/visible spectrophotometry were transferred to an NN-trained computer for modeling and prediction of output.…”