<title>Distance functions in dynamic integration of data mining techniques</title>

Abstract

One of the most important directions in the improvement of data mining and knowledge discovery is the integration of multiple data mining techniques. An integration method needs to be able either to evaluate and select the most appropriate data mining technique or to combine two or more techniques efficiently. A recent integration method for the dynamic integration of multiple data mining techniques is based on the assumption that each of the data mining techniques is the best one inside a certain subarea of the whole domain area. This method uses an instance-based learning approach to collect information about the competence areas of the mining techniques and applies a distance function to determine how close a new instance is to each instance of the training set. The nearest instance or instances are used to predict the performance of the data mining techniques. Because the quality of the integration depends heavily on the suitability of the used distance function, our goal is to analyze the characteristics of different distance functions. In this paper we investigate several distance functions as the very commonly used Euclidean distance function, the Heterogeneous Euclidean- Overlap Metric (HEOM), and the Heterogeneous Value Difference Metric (HVDM), among others. We analyze the effects of the use of different distance functions to the accuracy achieved by dynamic integration when the parameters describing datasets vary. We include also results of our experiments with different datasets which include both nominal and continuous attributes.