forestatrisk.predict package

forestatrisk.predict.interpolate_rho module

interpolate_rho(rho, input_raster, output_file='output/rho.tif', csize_orig=10, csize_new=1)

Resample rho values with interpolation.

This function resamples the spatial random effects (rho values) obtained from an iCAR model. It performs a cubicspline interpolation at a finer resolution and smoothens the rho values.

Parameters:

• rho – Original rho values estimates with the iCAR model.

• input_raster – Path to input raster defining the region.

• output_file – Path to output raster file with resampled rho values.

• csize_orig – Original size of the spatial cells (in km).

• csize_new – New size of the spatial cells for cubicspline interpolation (in km).

forestatrisk.predict.predict_raster module

predict_raster(model, _x_design_info, var_dir='data', input_forest_raster='data/forest.tif', output_file='predictions.tif', blk_rows=128, verbose=True)

Predict the spatial probability of deforestation from a statistical model.

This function predicts the spatial probability of deforestation from a statistical model. Computation are done by block and can be performed on large geographical areas.

Parameters:

• model – The model (glm, rf) to predict from. Must have a model.predict_proba() function.

• _x_design_info – Design matrix information from patsy.

• var_dir – Directory with rasters (.tif) of explicative variables.

• input_forest_raster – Path to forest raster (1 for forest).

• output_file – Name of the output raster file for predictions.

• blk_rows – If > 0, number of rows for computation by block.

• verbose – Logical. Whether to print messages or not. Default to True.

forestatrisk.predict.predict_raster_binomial_iCAR module

predict_binomial_iCAR(model, new_data, rhos)

Function to return the predictions of a model_binomial_iCAR model.

Function to return the predictions of a model_binomial_iCAR model for a new data-set. In this function, rho values for spatial cells are directly provided and not obtained from the model.

Parameters:

• model – The model_binomial_iCAR model to predict from.

• new_data – Pandas DataFrame including explicative variables.

• rhos – Spatial random effects for each observation (row) in new_data.

Returns:

Predictions (probabilities).

predict_raster_binomial_iCAR(model, var_dir='data', input_cell_raster='output/rho.tif', input_forest_raster='data/forest.tif', output_file='output/pred_binomial_iCAR.tif', blk_rows=128, verbose=True)

Predict the spatial probability of deforestation from a model.

This function predicts the spatial probability of deforestation from a model_binomial_iCAR model. Computation are done by block and can be performed on large geographical areas.

Parameters:

• model – The model_binomial_iCAR model to predict from.

• var_dir – Directory with rasters (.tif) of explicative variables.

• input_cell_raster – Path to raster of rho values for spatial cells.

• input_forest_raster – Path to forest raster (1 for forest).

• output_file – Name of the raster file to output the probability map.

• blk_rows – If > 0, number of rows for computation by block.

• verbose – Logical. Whether to print messages or not. Default to True.

forestatrisk.predict.wrast_rho module

wrast_rho(rho, input_raster, csize=10, output_file='output/rho_orig.tif')

Write rho values to GeoTIFF.

This function writes rho values (spatial random effects) to a GeoTIFF raster file.

Parameters:

• rho – original rho values estimates with the iCAR model.

• input_raster – path to input raster defining the region.

Csize:

size of the spatial cells (in km).

Output_file:

path to output raster file.