Easy Spatial Modeling with Random Forest
Prepares variable importance objects for spatial models
Custom print method for random forest models
Prints cross-validation results
Prints variable importance
Prints results of a Moran's I test
print_performance
Ranks spatial predictors
Rescales a numeric vector into a new range
Normality test of a numeric vector
Random forest models with Moran's I test of the residuals
Compares models via spatial cross-validation
Evaluates random forest models with spatial cross-validation
Contribution of each predictor to model transferability
Fits several random forest models on the same data
Fits spatial random forest models
Tuning of random forest hyperparameters via spatial cross-validation
RMSE and normalized RMSE
Finds optimal combinations of spatial predictors
Sequential introduction of spatial predictors into a model
Standard error of the mean of a numeric vector
Statistical mode of a vector
Suggest variable interactions and composite features for random forest...
Applies thinning to pairs of coordinates
Applies thinning to pairs of coordinates until reaching a given n
Variance Inflation Factor of a data frame
Transforms a distance matrix into a matrix of weights
Area under the ROC curve
Multicollinearity reduction via Pearson correlation
Multicollinearity reduction via Variance Inflation Factor
Defines a beowulf cluster
Generates case weights for binary data
Default distance thresholds to generate spatial predictors
Double centers a distance matrix
Removes redundant spatial predictors
Gets performance data frame from a cross-validated model
Gets the global importance data frame from a model
Gets the local importance data frame from a model
Gets Moran's I test of model residuals
Gets out-of-bag performance scores from a model
Gets model predictions
Gets model residuals
Gets data to allow custom plotting of response curves
Gets the spatial predictors of a spatial model
Checks if dependent variable is binary with values 1 and 0
Makes one training and one testing spatial folds
Makes training and testing spatial folds
Moran's Eigenvector Maps of a distance matrix
Moran's Eigenvector Maps for different distance thresholds
Moran's I test
Moran's I test on a numeric vector for different neighborhoods
Normality test of a numeric vector
Shows size of objects in the R environment
Optimization equation to select spatial predictors
Principal Components Analysis
PCA of a distance matrix over distance thresholds
Plots the results of a spatial cross-validation
Plots the variable importance of a model
Plots a Moran's I test of model residuals
Optimization plot of a selection of spatial predictors
Plot residuals diagnostics
Plots the response curves of a model.
Plots the response surfaces of a random forest model
Scatterplots of a training data frame
Moran's I plots of a training data frame
Plots a tuning object produced by rf_tuning()
Automatic generation and selection of spatial predictors for Random Forest models fitted to spatially structured data. Spatial predictors are constructed from a distance matrix among training samples using Moran's Eigenvector Maps (MEMs; Dray, Legendre, and Peres-Neto 2006 <DOI:10.1016/j.ecolmodel.2006.02.015>) or the RFsp approach (Hengl et al. <DOI:10.7717/peerj.5518>). These predictors are used alongside user-supplied explanatory variables in Random Forest models. The package provides functions for model fitting, multicollinearity reduction, interaction identification, hyperparameter tuning, evaluation via spatial cross-validation, and result visualization using partial dependence and interaction plots. Model fitting relies on the 'ranger' package (Wright and Ziegler 2017 <DOI:10.18637/jss.v077.i01>).