[R] randomForest - PartialPlot - reg

[Liaw, Andy]

In a partial dependence plot, only the relative scale, not absolute
scale, of the y-axis is meaningful.  I.e., you can compare the range of
the curves between partial dependence plots of two different variables,
but not the actual numbers on the axis.  The range is compressed
compared to the original data because of averaging.  For classification,
the function is computed in the logit scale, so it's not necessarily
positive.  High does mean higher probability for the target class.

Andy 

> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of Vijayan Padmanabhan
> Sent: Wednesday, September 22, 2010 11:47 PM
> To: r-help
> Subject: [R] randomForest - PartialPlot - reg
> 
> 
> Dear R Group
> I am not sure if this is the right forum to raise this query, 
> but i would 
> rather give it a try and aim for reaching the right person 
> who might be a 
> part of this group who can help.
> I have a query on interpretation of PartialPlot in package 
> randomForest. 
> In my earlier queries in this regard, I probably did not give 
> sufficient 
> explanation to elicit the intended details in the explanations being 
> provided.. Hence I am resending the query with examples and bit more 
> details.
> 
> In a scenario where a set of continuous variables vs a class 
> response is 
> being modeled by RF, say the iris example.. 
> using the following code, how do I interpret the partial plot that is 
> generated?
> 
> library(randomForest)
> data(iris)
> set.seed(543)
> iris.rf <- randomForest(Species~., iris)
> partialPlot(iris.rf, iris, Sepal.Length, "setosa")
> 
> How is the  y-axis values to be understood?
> 
> A straight forward Textual interpretation of the output  from 
> the experts 
> in this area, would help me understand this concept of 
> marginal effect 
> being plotted for the variable "Sepal.Length" on the 
> which.class=setosa.
> 
> Thanks for your help.
> 
> Regards
> Vijayan Padmanabhan
> 
> 
> "What is expressed without proof can be denied without proof" 
> - Euclide. 
> 
> 
> Can you avoid printing this?
> Think of the environment before printing the email.
> --------------------------------------------------------------
> -----------------
> Please visit us at www.itcportal.com
> **************************************************************
> ****************
> This Communication is for the exclusive use of the intended 
> recipient (s) and shall
> not attach any liability on the originator or ITC Ltd./its 
> Subsidiaries/its Group 
> Companies. If you are the addressee, the contents of this 
> email are intended for your 
> use only and it shall not be forwarded to any third party, 
> without first obtaining 
> written authorisation from the originator or ITC Ltd./its 
> Subsidiaries/its Group 
> Companies. It may contain information which is confidential 
> and legally privileged
> and the same shall not be used or dealt with by any third 
> party in any manner 
> whatsoever without the specific consent of ITC Ltd./its 
> Subsidiaries/its Group 
> Companies.
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 
Notice:  This e-mail message, together with any attachme...{{dropped:11}}