buildPredictor: Build a RFC and test its performances

In fernandoPalluzzi/tmsClassifier: Random Forest Classifiers for Transcranial Magnetic Stimulation Data

Description

Build a RFC from TMS regression parameters and test its classification performances.

Usage

buildPredictor(data, status = 10, vset = NULL, n = 10000, m = 3)

Arguments

data	A matrix or data.frame containing subjects as rows and TMS parameters as columns. A further column specifying subject diagnosis can be given to validate the classifier.
status	Numeric value specifying subject diagnosis (default status = 10).
vset	An optional data.frame of the same format of the input data. This will be used as external validation set. The validation set can be generated from input data, using tmsClassify.
n	Number of bootstrap sampled trees of the RFC (default n = 10000). Increasing n improves RFC classification performance, but will sensibly increase computational time.
m	Number of random variables for each tree split (default m = 3). The suggested value is the squared root of the total number of variables (the default value is set to 3 since 9 TMS parameter are expected by the default classifier).

Value

An object of class randomForest. An optional object of class performance is returned, if vset is not NULL.

References

Liaw A, Wiener M. Classification and Regression by randomForest (2002). R News, 2(3):18-22. https://doi.org/10.1023/A:1010933404324

See Also

randomForest

fernandoPalluzzi/tmsClassifier documentation built on Feb. 3, 2021, 12:31 p.m.