Fig. 3: Feature importance summaries.
From: Human movement and environmental barriers shape the emergence of dengue
Feature importance of gradient-boosted decision tree models for Mexico (A) and Brazil (D) using Shapley value summary plots. Colours indicate the relative values of each feature (rows) and their impact on model prediction (positive values = increased invasion risk). Features are ordered by mean impact on model output with the top variable conferring the most impact. Temp. temperature, Std dev standard deviation, env. veg. index environmental vegetation index. Timeseries plots and maps show if the contribution to the model-predicted invasion risk is greater for all environmental features or all mobility features for the year in which each municipality was invaded for Mexico (B, C) and Brazil (E, F). Source of Administrative boundaries: The Global Administrative Unit Layers (GAUL) dataset, implemented by FAO within the CountrySTAT and Agricultural Market Information System (AMIS) projects. Source data are provided as a Source Data file.
