Inductive learning: Algorithms and frontiers

摘要

Machine learning is a major subfield of artificial intelligence. It has been seen as a feasible way of avoiding the knowledge bottleneck problem in knowledge-based systems development. Research on machine learning has concentrated in the main on inductive learning. This paper surveys the current inductive learning research. The three typical inductive algorithms, AQ11, ID3 and HCV, are summarized with their main features being analyzed and three research frontiers, i.e., constructive learning, incremental learning and learning from data bases, in inductive learning are introduced.