Data as Ensembles of Records: Representation and Comparison

Abstract

Many collections of data do not come packaged in a form amenable to the ready application of machine learning techniques. Nevertheless, there has been only limited research on the problem of preparing raw data for learning, perhaps because widespread di#erences between domains make generalization di#cult. This paper focuses on one common class of raw data, in which the entities of interest actually comprise collections of (smaller pieces of) homologous data. We present a technique for processing such collections into high-dimensional vectors, suitable for the application of many learning algorithms including clustering, nearestneighbors, and boosting. We demonstrate the abilities of the method by using it to implement similarity metrics on two di#erent domains: natural images and measurements from ocean buoys in the Pacific. 1. Introduction A quick perusal of the UCI repository of machine learning data sets (Blake & Merz, 1999) reveals that the most frequently cite...