estimates.RandomForest: Report RandomForest object Estimates (effect sizes and...

View source: R/estimates.R

estimates.RandomForestR Documentation

Report RandomForest object Estimates (effect sizes and parameters)

Description

Report RandomForest object Estimates

Report RandomForest object Estimates

Usage

## S3 method for class 'RandomForest'
estimates(object, mc = FALSE)

## S3 method for class 'RandomForest'
estimates(object, mc = FALSE)

Arguments

object

a zeroinfl object

mc

Should model comparisons be performed? Currently not used

Value

One or more objects containing parameter estimates and effect sizes

Estimates for a RandomForest model. For classification, it will return the following:

  • OOB Accuracy in prediction: The total accuracy (correct classifications/incorrect classifications) for the OOB data

  • Variable importance: The difference in accuracy between the unshuffled and shuffled scores for the OOB data. Higher values indicate higher importance

For regression, it will return the following:

  • Quantiles of absolute value of OOB performance: This compares the OOB predicted values to the observed values, computing the difference between, and returns the quantiles.

  • Model R Squared: The squared correlation between the OOB predicted and the observed, just like in regular regression.

  • Variable importance: This is essentially the average difference in predictions between the original versus permutated values.


dustinfife/flexplot documentation built on June 12, 2025, 9:15 a.m.