RESUMO
Given the background of the use of Neural Networks in problems of apple juice classification, this paper aim at implementing a newly developed method in the field of machine learning: the Support Vector Machines (SVM). Therefore, a hybrid model that combines genetic algorithms and support vector machines is suggested in such a way that, when using SVM as a fitness function of the Genetic Algorithm (GA), the most representative variables for a specific classification problem can be selected.
Assuntos
Algoritmos , Bebidas , Frutas/química , Malus , Redes Neurais de Computação , Máquina de Vetores de Suporte , Bebidas/análise , Bebidas/classificaçãoRESUMO
Most of the plastics produced worldwide are finally disposed into the environment, most of them being one-use plastic packaging. Once released, plastics may undergone degradation through several agents, such as solar radiation, mechanical forces, and microbial action. Weathered plastics and microplastics (MPs) collected from the marine environment show considerable physical and chemical differences regarding their pristine counterparts; most notably on their surface, where spectrometric measurements are done. Hence, it is crucial to consider aging for their correct identification and quantification in environmental monitoring. Five of the most common polymers employed worldwide for packaging (LDPE, HDPE, PS, PP and PET) were weathered in a pilot-scale system simulating dry and marine conditions for more than 10 weeks. Aliquots were withdrawn periodically to monitor their weathering processes by means of infrared spectrometry and scanning electron microscopy; their spectra were compared and band ratios calculated. Results showed that an individual study of each polymer is necessary since degradation pathways and products depend on the polymer type. Moreover, including spectra of weathered polymers in the spectral libraries to obtain reliable identifications in microplastics pollution studies was critical.