Description Usage Arguments Value
A wrapper to produce a disease risk map from disease occurrence data and relevant environmental covariates following a standardised workflow. This function is intended for use in the ABRAID mapping platform, rather than for bespoke risk maps. As such the methods contained within are likely to change regularly and are not detailed here. At present they are based on the BRT ensemble approach in the Bhatt et al. (2013) dengue mapping paper.
1 2 3 4 5 6 7 8 9 10 11 12 |
mode |
A character vector of length one specifying type of model to be performed. E.g. 'Bhatt2013' to use the BRT ensemble approach from the Bhatt et al. (2013) dengue mapping paper. Supported modes can be found in 'data/abraid_modes.txt'. |
disease |
A numeric identifer for the disease being modelled. |
occurrence_path |
A character vector of length one specifying the file path of a csv file containing occurrence data. This file must contain the columns: 'Longitude', 'Latitude' (giving the coordinates of points), 'Weight' (giving the degree of weighting to assign to each occurrence record), 'Admin' (giving the admin level of the record - e.g. 1, 2 or 3 for polygons or -999 for points), 'GAUL' (the GAUL code corresponding to the admin unit for polygons, or NA for points) and 'Disease' a numeric identifer for the disease of the occurrence. |
extent_path |
A character vector of length one specifying the file path of a GeoTiff raster file giving the definitive extents of the disease. |
sample_bias_path |
A character vector of length one specifying the file path of a csv file containing occurrence data for other diseases. This file must contain the columns: 'Longitude', 'Latitude' (giving the coordinates of points), 'Admin' (giving the admin level of the record - e.g. 1, 2 or 3 for polygons or -999 for points), 'GAUL' (the GAUL code corresponding to the admin unit for polygons, or NA for points) and 'Disease' a numeric identifer for the disease of the occurrence. |
admin_path |
A named list representing locations of a set of GeoTiff raster file giving the GAUL codes for administrative regions at various levels. The names in the list should take the form "admin0" "admin1" etc. The values should be the file path for the various rasters. |
covariate_path |
A character vector specifying the file paths of GeoTiff raster files giving the environmetnal covariates used to predict risk of the disease. |
discrete |
A logical vector of the same length as |
verbose |
Whether to return messages from functions called in the script |
max_cpus |
The maximum number of CPUs to use when running the ensemble. At present |
load_seegSDM |
A parameterless function which can be used to load seegSDM and it's dependencies on each cluster node. |
parallel_flag |
Whether to perform parallelised run. |
Nothing is returned, but the following files are written to the working directory:
mean_prediction.asc |
a raster map giving the mean predicted disease risk |
prediction_uncertainty.asc |
a raster map giving uncertainty in the mean predicted disease risk (specifically the range of the 95% confidence interval) |
statistics.csv |
a csv giving the cross-validation statistics for each submodel in the ensemble |
relative_influence.csv |
a csv file summarizing the relative influence of each covaraite on the final model |
effect_plots.png |
a multi-panel plot of the marginal effect curves for each of the covariates used in the model |
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.