Abstract

A variety of techniques from statistics, signal processing, pattern recognition, machine learning, and neural networks have been proposed to understand data by discovering useful categories. However, research in data mining has not paid attention to the cognitive factors that make learned categories intelligible to human users. We show that one factor that influences the intelligibility of learned models is consistency with existing knowledge and describe a learning algorithm that creates concepts with this goal in mind. Introduction Knowledge-discovery in databases is a field whose goal is to extract usable models from a collection of data. Such models are expected to be accurate and are further expected to be intelligible to experts in the field. For example, knowledge acquired through such methods on a medical database might be published in scientific journals or written down as procedures to be followed in a health maintenance organization. While it is important that such knowled...

Citations