Project: Interactive construction, analysis and visualization of decision trees
Description
In this thesis the results of the ĕnal master project performed at the Eindhoven University of Technology, Department of Mathematics and Computer Science, Visualization group under the guidance of prof.dr.ir Jarke J. van Wijk are presented. The realized research has led to three products; this thesis, a prototype software system BaobabView and an extraction of this thesis is submitted to the annual VAST (Visual Analytics Science and Technology) conference.
We present a system for the interactive construction and analysis of decision trees that enables domain
experts to bring in domain speciĕc knowledge. We identify different user tasks and according
requirements, and develop a system incorporating a tight integration of visualization, interaction and
algorithmic support. Domain experts are supported in growing, pruning, optimizing and analysing
decision trees. Furthermore, we present a scalable decision tree visualization optimized for exploration.
We show the effectiveness of our approach by applying the methods to two use cases. e ĕrst case
illustrates the advantages of interactive construction, the second case demonstrates the effectiveness of
analysis of decision trees and exploration of the structure of the data.
Details
- Student
-
SEStef van den Elzen
- Supervisor
-
Jack van Wijk
- Link
- Thesis