Research On The Measurement Model Of Parallel Coordinates Visualization Technique

As the visualization techniques developing faster and faster, there have been more and more new kinds of visualization techniques. The appearance of a quantitative measurement model for visualization's application becomes one of the most imperative demands in the research field of visualization.This paper takes parallel coordinates as an example. After researching the parallel coordinates visualization technique and the theory of measurement model, a measurement model for visualization technique based on parallel coordinates was set up in this paper, which was carried out under Microsoft Visual C++ 6.0 environment.The research work of the paper is mainly embodied by the three aspects bellow:1. After collecting and arranging the resent research in and out of China, the paper summarized the achievements and limitations in visualization measurement field.2. The paper analyzed the merits and drawbacks of parallel coordinates visualization thoroughly, and abstracted the main deciding factors of parallel coordinates' application. After that, some index algorithms were designed and carried out, such as Intersection-Degree, Overlap-Degree and Distribution-Unbalance-Degree. The quantitative measurement model PVM was finally set up after weights of indexes were decided by AHP.3. The PVM model was carried out in C++. Dataset's parallel coordinates visualization can be estimated in the system, with a total score as well as values of other indexes. Also, the system carried out two applications of PVM: Choosing the best visualization and the index threshold control. There is also an experiment function designed in the system, which provides a scientific way of testing our model.The put forward of measurement model of visualization techniques based on parallel coordinates makes up for the shortcomings that people can only make subjective judgments of visualization's application, and motivates parallel coordinates visualization technique's application. Different visualization techniques have different features, and this paper focused on the research of parallel coordinates visualization. As to other kind of visualization techniques, there is still much room for researching.