Knowledge Discovery and Data Mining
Computers taught to discern patterns, detect anomalies and apply decision algorithms can help secure computer systems and find volcanoes on Venus
One of the most important ingredients of a scientist's work is the discovery of patterns in data. Yet the databases of modern science are frequently so immense that they preclude human analysis. In the past five years, investigators in the field of knowledge discovery and data mining have had notable successes in training pattern-detecting computers to do what people used to. In this article, Brodley, Lane and Stough recount some of these success stories, including the use of data mining to improve rotogravure printing, identify volcanoes in radar images of Venus and detect unwanted intruders on computer networks. They explain how some of the popular data-mining methods work, focusing particular attention on the method of decision trees, which produces "if-then" rules of the sort that humans can readily understand.
Go to Article