Inducer: a Rule Induction Workbench for Data Mining
Max Bramer · 2009
One of the key technologies of data mining is the automatic induction of classification rules from examples. A variety of methods have been proposed and many comparative studies of methods have been carried out. However the absence of a widely available common platform has made it difficult to carry out sufficiently in-depth studies to establish the superiority of one method over another. This paper describes Inducer, a rule induction workbench which aims to provide a common platform with a graphical user interface for analysing a variety of rule induction algorithms and related strategies with a range of datasets, without the need for any programming by the user. The principal features of Inducer are described with reference to two well-known rule induction algorithms. The paper concludes with an example of its use to compare these algorithms ' performance when there are missing data values.