Aspectos computacionais da modelagem de interação de características morfológicas entre plantas e polinizadores
Resumo
Resumo : A relação entre as características ou traços das espécies e suas interações em redes ecológicas tem sido amplamente discutida, especialmente no contexto de redes de polinização. A hipótese de que interações são mais prováveis quando os traços das espécies se complementam – por exemplo, quando a proporção entre a língua de um beija-flor e o formato de uma flor coincidem – tem motivado diversas investigações. No entanto, estimativas empíricas sobre a importância do pareamento de características variam significativamente entre diferentes tipos de redes ecológicas. Este trabalho explora essa variação com modelos de Aprendizagem de Máquina mais flexíveis, incluindo Florestas Aleatórias, Redes Neurais Profundas, Redes Neurais Convolucionais, Naive Bayes e KNN. Utilizando dados simulados e reais, investigamos a capacidade preditiva desses modelos para identificar interações planta-polinizador com base em traços Abstract : The relationship between species traits and their interactions in ecological networks has been widely discussed, especially in the context of pollination networks. The hypothesis that interactions are more likely when species traits complement each other, for example, when the proportion between a hummingbird’s tongue and the shape of a flower aligns, it has motivated various investigations. However, empirical estimates of the importance of trait-matching vary significantly across different types of ecological networks. This study explores this variation with more flexible Machine Learning (ML) models, including Random Forest, Gradient Boosting, Deep Neural Networks, Convolutional Neural Networks, Naive Bayes, and K-Nearest-Neighbors. Using simulated and real data, we investigate the predictive capacity of these models to identify plant-pollinator interactions based on traits