Mapping intra-urban malaria risk using high resolution satellite imagery: a case study of Dar es Salaam

Table 2 Summary of the average contributions of the significant predictor variables using a Boosted Regression Trees (BRT) model

Variable	Average relative contribution (%)
Percentage dense/riverine vegetation	29.94
Percentage built-up	26.85
Distance to inland water (m)	8.98
Altitude	8.34
Wetness Index (CTI)	6.61
NDVI	5.44

BRT model developed with cross-validation over 25 bootstraps. Average relative contribution refers to the influence of each variable to the BRT model calculated as the proportion of times that a variable was selected for splitting, weighted by the squared improvement to the model as a result of each split [73]. This was then averaged over the 25 iterations of the BRT model run. This has added to the footnote of Table 2. Variables with zero influence or relative contribution <1 % were dropped from the analysis. Effect of other land cover classes was controlled for

ISSN: 1476-072X