The RIPPER Algorithm

Machine Learning- Understanding of classification rules

Classification Rules

Classification rules represent knowledge in the form of logical if-else statements that assign a class to unlabeled examples. An ‘antecedent’ and a ‘consequent’ are the terms for them. This form a statement that says ‘if this happens, then that happens’.

The earlier rule learning algorithms (Separate and conquer, and The 1R algorithm) have some problems like slow performance for an increasing number of datasets, and prone to being inaccurate on noisy data.

Johannes Furnkranz and Gerhard Widmer in 1994 proposed a solution towards solving these problems. Their incremental reduced error pruning algorithm (IREP) uses a combination of pre-pruning and post-pruning methods that grow very complex rules and prune them before separating the instance from the complete dataset.

The RIPPER (repeated incremental pruning to produce error reduction)algorithm is introduced by W. Cohen in 1995, which improved upon IREP to generate rules that match or exceed the performance of decision trees. Having evolved from several iterations of the rule learning algorithm, the RIPPER algorithm can be understood in a three-step process.

• Grow
• Prune
• Optimize