Orange Forum • View topic - Missing documents on Tree classifier

Missing documents on Tree classifier

Report bugs (or imagined bugs).
(Archived/read-only, please use our ticketing system for reporting bugs and their discussion.)
Forum rules
Archived/read-only, please use our ticketing system for reporting bugs and their discussion.

Missing documents on Tree classifier

Postby aurora1625 » Sat Sep 04, 2010 7:44

I am learning Python and Orange by reading your document and planning to build my own tree-classifier.

I have to say the first part of the document is quite extensive and specific and then I got this:
Well, this might have been an overkill...
Splitting criteria
Example splitter
Flags and similar
... also mention nodeLearner and descender
Programming your own tree learner skeleton
Node classifier
Same majority pruning
Post pruning
... show a series of trees
Defining your own components
Various tricks
Storing testing examples
... show how to remember misclassifications etc.

Replacing node classifiers
... replacing with something else, but based on learning examples

When can this document be filled, I so want to learn this part.

Postby Janez » Tue Sep 14, 2010 21:54

The page has been in such a state for a long time and may stay so for a while: you said that this document is quite extensive, but the part that is missing is probably even much larger.

But that's what forum is for. What exactly would you want to do?

If you just want to induce a tree classifier, you'd better look here: It's still long, but the part you're interested in is in the beginning.

If you want to program a tree inducer, start here: It's not about trees but about Bayesian classifier; the principle is the same, although the algorithm and the data structures are of course more complex.

Only if you want to change some components of the existing inducer, e.g. try a new method for treating missing data, you actually need the missing documentation. Ask here instead, tell us what you need and we'll help.

Return to Bugs