I am a new beginner who recently started using the Random forest model in R. I ran an analysis on my data and received the following results. I just want to understand what does the "squared residuals mean" mean in the results I got? Is that the difference between the actual value and the predicted value by the model? If the mean of residuals is this high (62.20) then why the % var explained is also high? Doesn't the "% var explained" mean the amount of "out-of-bag" data correctly predicted by the model? If 78% of data was correctly predicted, then how come there is such a huge difference between the predicted and target variables? Thank you so much!

What is the range of `hw.area.10`

?

Hi nigrahamuk,

Sorry for the delay in answering. The range of hw.area.10 is very broad. It starts from 82 and goes up to 180.

Then consider. 62 is not mean of residuals. Its mean of squared residuals. Your squared range is on the order of 6724 to 32400

But does this " mean of residuals" shows the difference between my predicted values and my target values? Does it mean that some of my predicted values have a great difference from my target values? thank you!

Define 'great' difference ?

You are out by a mean of +/- 8 is that great ?

Again my apologies because I am pretty new to these concepts. I understand that the mean of squared residuals is 62. So does that mean that some of my predicted variables are +/- 8 different than my target variables? Can the difference also be more than +/- 8 ? because this value is just a mean? thank you!

it means that the mean of the difference is on that order ~8, yes.

the difference is almost certainly more in a good number of cases (and less ) such is the nature of averages.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.