Dimensionality Reduction in ILP: A Call to Arms
Johannes Fürnkranz · 1997
The recent uprise of Knowledge Discovery in Databases (KDD) has underlined the need for machine learning algorithms to be able to tackle largescale applications that are currently beyond their scope. One way to address this problem is to use techniques for reducing the dimensionality of the learning problem by reducing the hypothesis space and/or reducing the example space. While research in machine learning has devoted considerable attention to such techniques, they have so far been neglected in ILP research. The purpose of this paper is to motivate research in this area and to present some results on windowing techniques. 1 Introduction One of the most often heard prejudices against ILP algorithms is that they are only applicable to toy problems and will not scale up to applications of significant size. While it is our firm belief that the order of magnitude of this unspecified "significant size" is monotonicly increasing in order to keep the argument alive, it is nevertheless indis...