Using Case-Based Reasoning in Interpreting Unsupervised Inductive Learning Results

The objective of this work is to interpret inductive results obtained by the unsupervised learning method OSHAM. We briefly introduce the learning process of OSHAM, that extracts concept hierarchies from unlabelled data, based on a representation combining the classical, prototype and exemplar views on concepts. The interpretive process is considered as an intrinsic part in OSHAM and is carried out by a combination of case-based reasoning with matching approaches in inductive learning. An experimental comparative study of some learning methods in terms of knowledge description and prediction is given.

• Type: Conference
• Year: 1997
• Where: IJCAI
• Authors: Tu Bao Ho, Chi Main Luong
• Keywords: IJCAI 1997, IJCAI 2007, Learning Methods, Learning Process, Unsupervised Learning Method