Table 4 The prediction error obtained from the fivefold cross-validation under six combinations of the two tuning parameters.

From: Functional random forests for curve response

\(L^\text{{{a}}}\)

\(mtry^\text{{{b}}}\)

10

20

35

Adaptive L

119.5

118.9

119.4

Fixed \(L = 4\)

119.7

119.3

118.9

  1. \(^\text{{{a}}}\)L is the number of functional PCs to be retained in Eq. (3).
  2. \(^\text{{{b}}}\)mtry is the size of the subset of the predictors that are considered at each split.
Back to article page