IMPLEMENTATION OF PATTERN RECOGNITION TECHNIQUES (SUPPORT VECTOR MACHINES - LEAST SQUARE) IN SELECTION OF PARAMETERS USED IN METABOLOMIC SYSTEMS IMPLEMENTACIÓN DE TÉCNICAS DE RECONOCIMIENTO DE PATRONES (LEAST SQUARE SUPPORT VECTOR MACHINES) EN PROCESOS DE SELECCIÓN DE PARÁMETROS CARACTERÍSTICOS APLICADOS A SISTEMAS METABOLÓMICOS
This paper presents a metodologia que involved, multivariate analysis techniques and pre-processing stage in order to determine characteristic metabolites in a given spectrum. This novel approach allowed us to determine that certain metabolites are modified by the addition of different concentrations of about feature SVM LS-NMR data. Validating processes also achieved as peak alignment, normalization, baseline correction and analysis multienergy in metabolomic data in olive oils and pure and mixed with hazelnut alterations 2%, 5%, 10%, 20% and 30% .
