A Multi-dimensional Data Plotting Model Using Hybrid Parallel Coordinates Based Visualization

Abstract

Graphics based visualization system had shown ever expanding growth in terms of their applications and devices supportability. Aim is towards effective plotting of information. But if the information is increase with limiting plotting ranges it requires another types of geometric projection scheme termed as parallel coordinates. These are used for representing multidimensional data on to a single graphs or image. Traditionally these systems are using various plotting and transformation mechanism which defines shapes and its boundaries such as lines, poly-lines, planes, hyper planes, curves etc. These approaches are normally stuck into problems related to local minima, cross over plotting, outliers and dimensionality reductions, by which plotting of multiple information on single points make the visualization disrupt with unidentified noises signal. To overcome the above issues this work suggested some refinements and improvements strategies. The paper proposes a novel hybrid parallel coordinate based multidimensional data plotting. The systematic working with improved curve based plotting remaining problems in existing works can be easily resolved. Apart from basic functionality, the future implementation of tools supports data plotting from huge data sets with some more mechanism like filtering, brushing, specialized curves, colouring etc. At the analytical level of results evaluation, approach is showing its strong presence.

 

Index Terms: Visualization System, Geometric Projection, Parallel Coordinates (PC), Filtering, Curves, Brushing, Colouring, Hybrid Approach;