Application of Computational Intelligence Techniques for Forecasting Problematic Wine Fermentations Using Data from Classical Chemical Measurements
The early forecasting of normal and problematic wine fermentations is one of the main problems of winemaking processes, due to its significant impacts in wine quality and utility. In Chile this is a critical problem because it is one of the top ten wine-producing countries. In this chapter, we review the computational intelligence methods that have been applied to solve this problem. Both methods studied, support vector machines and artificial neural networks, show excellent results with respect to the overall prediction error for different training/testing/validation percentages, different time cutoffs, and several parameter configurations. These results are of great importance for wine production because they are based only on measurement of classical chemical variables and they confirm that computational intelligence methods are a useful tool to the winemakers in order to correct in time a potential problem in the fermentation process.
KeywordsSupport vector machines Neural networks Wine fermentations Classical chemical variables
Research supported by grants: FONDECYT 1120679 and Conicyt PIA/Basal FB0821.
- Bishop CM (1996) Neural networks for pattern recognition. Oxford University Press, OxfordGoogle Scholar
- Bishop CM (2006) Pattern recognition and machine learning. Springer, New YorkGoogle Scholar
- Bisson L, Butzke C (2000) Diagnosis and rectification of stuck and sluggish fermentations. Am J Enol Vitic 51(2):168–177Google Scholar
- Demyanov S, Bailey J, Ramamohanarao K, Leckie C (2012) AIC and BIC based approaches for SVM parameter value estimation with RBK kernels. JMLR W&CP 25:97–112Google Scholar
- Engelbrecht AP (2007) Computational intelligence: an introduction. 2nd ed. John Wiley & Sons, West Sussex, EnglandGoogle Scholar
- Executive Report Chilean Wine Production (2011–2016) Servicio Agrícola y Ganadero de ChileGoogle Scholar
- Hernández G, Leon R, Urtubia A (2017) Application of neural networks for the early prediction of problematic wine fermentations using data from the classical chemical measurements, submitted to Journal of Biotechnology, (in press)Google Scholar
- Kruse R, et al (2013) Computational intelligence: a methodological introduction. Springer-Verlag, London, EnglandGoogle Scholar
- Pszczólkowski P, Carriles P, Cumsille M, Maklouf M (2001) Reflexiones sobre la madurez de cosecha y las condiciones de vinificación, con relación a la Problemática de fermentaciones alcohólicas lentas y/o paralizante en Chile, Facultad de Agronomía, Pontificia Universidad Católica de ChileGoogle Scholar
- Ripley BD (2008) Pattern recognition and neural networks. Cambridge University Press, Cambridge, USAGoogle Scholar
- Sánchez D (2003) Advanced support vector machines and kernel methods. NeurocomputingGoogle Scholar
- Scholkopf B, Smola AJ (2002) Learning with kernels: support vector machines, regularization, optimization, and beyond. The MIT Press, Cambridge, USAGoogle Scholar