[R] sign of the y axis in partialPlot for randomForest regression

[Cédric Mondy]

Hi everybody,

I used randomForest to regress invertebrates abundances in least impaired
river reaches from some environmental parameters. Then I used these models
to predict invertebrates abundances in impaired reaches.

Now I would like to model the deviation (observation - prediction) with a
set of chemical parameters to see if the deviations from predictions could
be explained with water chemistry.

I did built the model, and I used partialPlot to  depict the patterns
between individual water chemistry parameters and deviation from prediction.

I know that the range and the values indicated on the y-axis do not
correspond to the 'raw' deviation. However, since my deviations could be
positive (i.e. greater abundance than expected) or negative (i.e. lower
abundance than expected), I would like to know if the signs of the y-axis do
correspond to those of my deviation values: does a negative value in the
y-axis really correspond to a negative value of deviation.


Thanks for your help

Cédric 

--
View this message in context: http://r.789695.n4.nabble.com/sign-of-the-y-axis-in-partialPlot-for-randomForest-regression-tp3755583p3755583.html
Sent from the R help mailing list archive at Nabble.com.