A methodology applicable to building a classifier of pavement roughness measurement methods and devices

Abstract

The paper presents a methodological approach used by the authors in the process of design of a knowledge base for the domain related to measurement of different pavement characteristics. As the main representation formalism a rule is considered. Individual steps taken in the process of knowledge discovery in data are discussed. The process is demonstrated using limited sample data related to pavement roughness measurements processed by inductive learning algorithms. The paper emphasizes procedural aspects of the design process rather than presentation of complete results that are not fully available yet because of the early stage of the research project being in the background.