Workflow 4

Workflow 4: Application of machine learning for ecological analysis

This analytical workflow studies the application of machine learning models for ecological analysis. This analytical workflow explores a range of machine learning algorithms with a regression approach, such as eXtreme Gradient Boosting (XGBoost), multiple linear regression, neural networks, random forests and support vector machines, that can be flexibly applied to diverse ecological analyses. The workflow incorporates SHAP (SHapley Additive exPlanations) analysis to interpret individual predictions and determine the significance of global features by attributing a model’s output to its input features. This workflow is designed to work with any numerical dataset and with different numbers of columns.