bmaPredict: bmaPredict

In DataScienceSalon/Bayesian-Regression: Bayesian Linear Regression and Default Priors

Description

bmaPredict Performs prediction using a single model and estimator

Usage

bmaPredict(model, estimator, yX = NULL, trial = NULL, prediction = FALSE,
  rvp = FALSE, pe = FALSE, predObj = FALSE)

Arguments

model	BAS.LM object
estimator	String indicating which estimator to use. Options are 'BMA', 'BPM', 'HPM', & 'MPM'
yX	Optional data frame containing test data
trial	Numeric indicator of the trial number when repeated trials are being conducted.
prediction	Logical value to indicate whether the observed design matrix used in fitting or the newdata will be used for estimating the mean or for making predictions. The default is FALSE for estimation of the mean.
rvp	Logical indicating whether to return residuals vs predicted (fitted) data frame
pe	Logical indicating whether to return paramater estimates
predObj	Logical. If TRUE, the object from the predict function is returned.

Details

Performs predictions on a single model and estimator and returns MSE and optionally, residual vs fitted, parameter estimates and prediction distributions on existing training data or new test data.

Value

mse frame containing MSE measures for each model

Author(s)

John James, [e-mail]

See Also

Other BMA functions: bestPredictions, bmaAnalysis, bmaComplexity, bmaEvaluation, bmaPDC, bmaPIPPlots, bmaPIP, bmaPerformanceReport, bmaPerformance, bmaPredictModels, bma

DataScienceSalon/Bayesian-Regression documentation built on May 29, 2019, 12:06 a.m.