random forest predicted variance

I was looking at the predictions of my random forest model and saw the predicted variance that was outputted. I noticed that some of these predicted values have very large variance values. I was just wondering if anyone knew how this variance value was calculated. Is it just a normal variance formula with the sum of the squared difference between predicted value each tree and average predicted value?

Hi @Haseeb23

The variance column contains the unbiased variance of the predictions of the individual trees.
It is calculated using org.apache.commons.math.stat.descriptive.moment.Variance.
See https://commons.apache.org/proper/commons-math/javadocs/api-1.0/org/apache/commons/math/stat/descriptive/moment/Variance.html for the Apache documentation.

3 Likes

This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.