g

[

g,

,

]

R function for the random forest algorithm is in the

Forest

package.

The

R

function

is

also

called

Forest. The random forest algorithm can be used for both

tion and regression analysis. A random forest model is

ed using the following code,

randomForest(formula,data,⋯)

e 3.49(a) shows the ROC curve of a random forest model for the

ncer data.

andom forest algorithm can also analyse non-numerical data. This

e it is developed based on decision tree algorithms. The factor Xa

data was used for the demonstration. Figure 3.49(b) shows the

ve for this random forest model constructed for the data, where

ing process was used to encode the peptides.

dition to randomForest, the other package named as party

be used for random forest data analysis. The function is named as

t. The format of using this package for constructing a random

del using party is shown below,

cforest (formula,data,control)

e 3.49(c) shows the ROC curve of a random forest model

ed using the party package for the factor Xa protease cleavage

ain, no encoding process was used for the amino acids in the

dvantage of the party package is that it can generate a tree for

ation and visualisation. For instance, Figure 3.50 shows a tree

d by the party package for the breast cancer data set. Using this

possible to find which variables play the significant roles in the

mour classification problem, hence providing useful knowledge

t cancer diagnosis. The tree also shows the significance of each

i.e., a p value associated with a variable. For instance, the

X24 had the least p value and hence was the most significant

for breast cancer diagnosis based on a model constructed for this