Thumbnail Image

Modular System for Comparison of Navigation Algorithms in Visual Data Exploration

Jusuf, Patricia
Navigation of visualization process involves large, complex, unclassified data sets and multidimensional, nonlinear, discontinuous mapping functions. Thus, getting a desirable data from those data sets is usually a painful process. It is then important to know what is the best technique to do the navigation process. Many algorithms have been developed to improve the navigation process. But since each of those algorithms has its own approach and test data sets, a system that can compare the performance of those algorithms is needed. This thesis is about a modular system that is designed especially for testing and comparing those algorithms with various data sets. The system is divided into six components. Each component can have several different types and can be easily taken off from the system and substituted with other components. The system uses network so that many users can access the system and giving feedback at the same time. This can help speeding up the navigation process. XML template is used to assign values to test data sets instead of having several test data set files. We will compare the system proposed with several test data sets: OpenGL standard objects, real world objects, and scientific data sets.