Overhaul of the prediction interface for BART and BCF models

DESCRIPTION

@@ -29,7 +29,7 @@ Description: Flexible stochastic tree ensemble software.
 License: MIT + file LICENSE
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.3.2
+RoxygenNote: 7.3.3
 LinkingTo: 
     cpp11, BH
 Suggests: 

NAMESPACE

@@ -10,6 +10,10 @@ export(calibrateInverseGammaErrorVariance)
 export(computeForestLeafIndices)
 export(computeForestLeafVariances)
 export(computeForestMaxLeafIndex)
+export(compute_bart_posterior_interval)
+export(compute_bcf_posterior_interval)
+export(compute_contrast_bart_model)
+export(compute_contrast_bcf_model)
 export(convertPreprocessorToJson)
 export(createBARTModelFromCombinedJson)
 export(createBARTModelFromCombinedJsonString)
@@ -60,6 +64,8 @@ export(rootResetRandomEffectsModel)
 export(rootResetRandomEffectsTracker)
 export(sampleGlobalErrorVarianceOneIteration)
 export(sampleLeafVarianceOneIteration)
+export(sample_bart_posterior_predictive)
+export(sample_bcf_posterior_predictive)
 export(sample_without_replacement)
 export(saveBARTModelToJson)
 export(saveBARTModelToJsonFile)
@@ -77,6 +83,7 @@ importFrom(stats,pnorm)
 importFrom(stats,predict)
 importFrom(stats,qgamma)
 importFrom(stats,qnorm)
+importFrom(stats,quantile)
 importFrom(stats,resid)
 importFrom(stats,rnorm)
 importFrom(stats,runif)

NEWS.md

@@ -2,13 +2,30 @@
 
 ## New Features
 
+* Support for multithreading in various elements of the GFR and MCMC algorithms ([#182](https://github.com/StochasticTree/stochtree/pull/182))
 * Support for binary outcomes in BART and BCF with a probit link ([#164](https://github.com/StochasticTree/stochtree/pull/164))
+* Enable "restricted sweep" of tree algorithms over a handful of trees ([#173](https://github.com/StochasticTree/stochtree/pull/173))
+* Support for multivariate treatment in R ([#183](https://github.com/StochasticTree/stochtree/pull/183))
+* Enable modification of dataset variables (weights, etc...) via low-level interface ([#194](https://github.com/StochasticTree/stochtree/pull/194))
+
+## Computational Improvements
+
+* Modified default random effects initialization ([#190](https://github.com/StochasticTree/stochtree/pull/190))
+* Avoid double prediction on training set ([#178](https://github.com/StochasticTree/stochtree/pull/178))
 
 ## Bug Fixes
 
 * Fixed indexing bug in cleanup of grow-from-root (GFR) samples in BART and BCF models
 * Avoid using covariate preprocessor in `computeForestLeafIndices` function when a `ForestSamples` object is provided (rather than a `bartmodel` or `bcfmodel` object)
 
+## Other Changes
+
+* Standardized naming conventions for out of sample data in prediction and posterior computation routines (we raise warnings when data are passed through `y`, `X`, `Z`, etc... arguments)
+    * Covariates / features are always referred to as "covariates" rather than "X"
+    * Treatment is referred to as "treatment" rather than "Z"
+    * Propensity scores are referred to as "propensity" rather than "pi_X"
+    * Outcomes are referred to as "outcome" rather than "Y"
+
 # stochtree 0.1.1
 
 * Fixed initialization bug in several R package code examples for random effects models