Is there a well-designed, maintained decision tree learning library for Java?

Question

I need a decision tree learning library for Java. I've looked at both jaDTi, and Weka, but neither is up to contemporary standards of library design.

For example, both still use non-generic Vector objects everywhere, and neither makes it easy to supply training data programmatically, they both expect data to be loaded from a file or a database.

Answer 1

I would suggest you take a look at JBoost.

Answer 2

Apache Mahout has a decision tree implementation based on random forests. It is Hadoop-based however, not plain Java.