gbm.step.sd: Function to assess optimal no of boosting trees using k-fold...

In SimonDedman/gbm.auto: Automated Boosted Regression Tree Modelling and Mapping Suite

Function to assess optimal no of boosting trees using k-fold cross validation

Description

SD fork of dismo's gbm.step to add evaluation metrics like d.squared and rmse. J. Leathwick and J. Elith - 19th September 2005, version 2.9. Function to assess optimal no of boosting trees using k-fold cross validation. Implements the cross-validation procedure described on page 215 of Hastie T, Tibshirani R, Friedman JH (2001) The Elements of Statistical Learning: Data Mining, Inference, and Prediction Springer-Verlag, New York.

Usage

gbm.step.sd(
  data,
  gbm.x,
  gbm.y,
  offset = NULL,
  fold.vector = NULL,
  tree.complexity = 1,
  learning.rate = 0.01,
  bag.fraction = 0.75,
  site.weights = rep(1, nrow(data)),
  var.monotone = rep(0, length(gbm.x)),
  n.folds = 10,
  prev.stratify = TRUE,
  family = "bernoulli",
  n.trees = 50,
  step.size = n.trees,
  max.trees = 10000,
  tolerance.method = "auto",
  tolerance = 0.001,
  plot.main = TRUE,
  plot.folds = FALSE,
  verbose = TRUE,
  silent = FALSE,
  keep.fold.models = FALSE,
  keep.fold.vector = FALSE,
  keep.fold.fit = FALSE,
  ...
)

Arguments

data	The input dataframe.
gbm.x	The predictors.
gbm.y	The response.
offset	Allows an offset to be specified.
fold.vector	Allows a fold vector to be read in for CV with offsets,.
tree.complexity	Sets the complexity of individual trees.
learning.rate	Sets the weight applied to inidivudal trees.
bag.fraction	Sets the proportion of observations used in selecting variables.
site.weights	Allows varying weighting for sites.
var.monotone	Restricts responses to individual predictors to monotone.
n.folds	Number of folds.
prev.stratify	Prevalence stratify the folds - only for p/a data.
family	Family - bernoulli (=binomial), poisson, laplace or gaussian.
n.trees	Number of initial trees to fit.
step.size	Numbers of trees to add at each cycle.
max.trees	Max number of trees to fit before stopping.
tolerance.method	Method to use in deciding to stop - "fixed" or "auto".
tolerance	Tolerance value to use - if method == fixed is absolute, if auto is multiplier * total mean deviance.
plot.main	Plot hold-out deviance curve.
plot.folds	Plot the individual folds as well.
verbose	Control amount of screen reporting.
silent	To allow running with no output for simplifying model).
keep.fold.models	Keep the fold models from cross valiation.
keep.fold.vector	Allows the vector defining fold membership to be kept.
keep.fold.fit	Allows the predicted values for observations from CV to be kept.
...	Allows for any additional plotting parameters.

Details

Divides the data into 10 subsets, with stratification by prevalence if required for pa data then fits a gbm model of increasing complexity along the sequence from n.trees to n.trees + (n.steps * step.size) calculating the residual deviance at each step along the way after each fold processed, calculates the average holdout residual deviance and its standard error then identifies the optimal number of trees as that at which the holdout deviance is minimised and fits a model with this number of trees, returning it as a gbm model along with additional information from the cv selection process.

D squared is 1 - (cv.dev / total.deviance). Abeare thesis: For each of the fitted models, the pseudo-R2, or D2, or Explained Deviance, was calculated for comparison, where: D2 = 1 – (residual deviance/total deviance).

requires gbm library from Cran requires roc and calibration scripts of J Elith requires calc.deviance script of J Elith/J Leathwick

Value

GBM models using gbm as the engine.

SimonDedman/gbm.auto documentation built on Oct. 9, 2024, 8:57 p.m.