|
|
Rule Learner is a rule induction program that covers the training examples with replacement and learns agnostic models. An agnostic model is one that may abstain from making a predicion on a given query.
RL has a number of parameters that allow the user to set an approptiate learning bias, using knowledge knowledge about the learning problem.
For feature selection, RL has a randomized greedy wrapper that runs a specified number of times and summarizes results.
To find important attributes, RL tests the sensitivity of learned models to removing subsets of attributes.
|
|
|
Maintainer
|
|
Project Members
|
|
Past Members
|