Information Visualization: Which Is the Most Appropriate Technique to Represent Data Mining Results?

Abstract

Data mining (DM) is a research area that has been contributing in the search for implicit knowledge that can give support to decision making. However, the analysis of the results obtained by using the techniques of DM is not an easy task. Such difficulty can be minimized, by using techniques of information visualization. But the issue is: which will be the most appropriate technique for each type of result to be analyzed? To help in the choice of techniques, this paper presents an evaluation of techniques for information visualization. Such task was accomplished by using the method known as analysis of characteristics, having as result the analysis on the effectiveness when using techniques of geometric and iconographical information visualization, mainly in relation to the results obtained with the application of k-means algorithm.