Fig. 2
From: ‘Forest malaria’ in Myanmar? Tracking transmission landscapes in a diversity of environments
Schematic of the statistical plan. Hierarchical ascendant clustering (HAC) on principal components analysis (PCA) built up landscapes and climates based on land use land cover (LULC) or rainfall and temperature data. Conditional inference trees (CIT) and conditional random forests (CRF) realized at region and township scales explored associations between environmental factors and malaria incidence profiles. Eco-epidemiological zones were constructed based on climates and distance to landscapes significantly associated with Plasmodium falciparum and Plasmodium vivax incidence profiles in CIT and CRF. METF Malaria elimination task force