MDR (version 1.2)

summary.mdr: Summarizing the results of an MDR model

Description

summary method for class 'mdr', after fitting with mdr.cv or mdr.3WS

Usage

"summary"(object,...)

Arguments

object

an object of class 'mdr', a result of a call to either mdr.cv or mdr.3WS

...

additional arguments

Value

a table, with columns for level of interaction, bests MDR models for each level (including overall best model), and accuracy results. Accuracy results depend on the validation method for object.

Level

level of interaction

Best Models

best MDR by level

Classification Accuracy

average classification accuracy (percent) calculated from the training sets; for mdr.cv

Prediction Accuracy

average prediction accuracy (percent) calculated from the testing sets; for mdr.cv

Cross-Validation Consistency

the number for times a model was chosen as 'best' out of k, for k-fold cross-validation; for mdr.cv

Training Accuracy

classification accuracy (percent) calculated from the training set; for mdr.3WS

Testing Accuracy

classification accuracy (percent) calculated from the testing set; for mdr.3WS

Validation Accuracy

classification accuracy (percent) calculated from the validation set; for mdr.3WS

...

See Also

mdr.cv, mdr.3WS, predict.mdr, plot.mdr

Examples

#load test data
data(mdr1) #consider a subset with the response and the first 10 predictors

fit1<-mdr.3WS(data=mdr1[,1:11],K=2) #fit mdr model with 3WS
summary(fit1) #summarizes results of the fit

fit2<-mdr.cv(data=mdr1[,1:11],K=2,cv=5) #fit mdr model with 5-fold CV
summary(fit2)