[R] interperting a regression tree

[Bill.Venables at csiro.au]

If you have fitted a tree model, the diagram (which did not survive the
mailing list, of course) would have been a decision tree, not a
dendrogram.

In a decision tree, the internal nodes refer to decisions to be made as
to whether to proceed to the left or to the right at that stage.  This
is based on the value of a selected predictor variable.  If tm is the
tree model, you would normally display it using

plot(tm)
text(tm)

Each internal node is then labelled with a message line "age < 25.5",
which shows that at that point the decision is made on the variable,
'age'.  If the age of the case is less than 25.5 choose the left branch,
otherwise the right branch.

When you arrive at a terminal node, bingo!  You get a prize.

Bill Venables 



-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
On Behalf Of Sharma, Manju
Sent: Thursday, 20 March 2008 8:54 PM
To: r-help at r-project.org
Subject: [R] interperting a regression tree

hello
i am trying to decipher a dendrogram i have from performing a 'tree'.
(attached file) 
my response variable is factored - low, medium and high threat and the 8
explanatory variables are numeric values. 
 
i could do with some help to understand what the values <0.25, <0.5 and
<0.75 are on the tree branches? how does one interpret this??? 
 
cheers,
mann
 
 
 
Manju Sharma
Ashoka Trust for Research in Ecology and the Environment,
Banaglore, India