randomUniformForest (version 1.1.6)

model.stats: Common statistics for a vector (or factor) of predictions and a vector (or factor) of responses

Description

Given a vector of predictions and a vector of responses, provide some statistics and plots like AUC, AUPR, confusion matrix, F1-score, geometric mean, residuals, mean squared and mean absolute error.

Usage

model.stats(predictions, responses, regression = FALSE, OOB = FALSE, plotting = TRUE)

Arguments

predictions

a vector (or factor, if classification) of predictions.

responses

a vector (or factor, if classification) of responses of the same length than 'predictions'.

regression

if FALSE, considered arguments are treated as a classification task.

OOB

if TRUE, expects 'prediction' to be an object of class randomUniformForest (with option 'OOB' enabled) in order to assess OOB predictions.

plotting

if TRUE, displays graphics. Set it to FALSE in the case of a regression with large datasets.

Value

print and plot metrics.

Examples

# NOT RUN {
## not run
## Classification : synthetic data

# set.seed(2014)
# n = 1000
# p = 100 
# X = simulationData(n, p)
# X = fillVariablesNames(X)
# epsilon1 = runif(n,-1,1)
# epsilon2 = runif(n,-1,1)
# rule = 2*(X[,1]*X[,2] + X[,3]*X[,4]) + epsilon1*X[,5] + epsilon2*X[,6]
# Y = as.factor(ifelse(rule > mean(rule), "+","-")) 

# training and test sets

# train_test = init_values(X, Y, sample.size = 1/2)
# X1 = train_test$xtrain
# Y1 = train_test$ytrain
# X2 = train_test$xtest
# Y2 = train_test$ytest

# train model
# synth.ruf = randomUniformForest(X1, as.factor(Y1))

# evaluates OOB predictions
# statsOOB.pred.synth.ruf = model.stats(synth.ruf, as.factor(Y1), OOB = TRUE)

# predict 
# pred.synth.ruf = predict(synth.ruf, X2)

# statistics : produces also two plots
# stats.pred.synth.ruf = model.stats(pred.synth.ruf, as.factor(Y2))

# or, trick, do all in two lines
# synth.ruf = randomUniformForest(X1, as.factor(Y1), xtest = X2, ytest = as.factor(Y2))
# stats.pred.synth.ruf = model.stats(synth.ruf, as.factor(Y2))

## regression : synthetic data
# Y = rule
# Y1 = Y[train_test$train_idx]
# Y2 = Y[train_test$test_idx]

# synth.ruf = randomUniformForest(X1, Y1)
# statsOOB.pred.synth.ruf = model.stats(synth.ruf, Y1, OOB = TRUE, regression = TRUE)
# pred.synth.ruf = predict(synth.ruf, X2)
# stats.pred.synth.ruf = model.stats(pred.synth.ruf, Y2, regression = TRUE)
# }