Year: 2015
Type: article
Abstract: "A picture is worth a thousand words.” This idiom is true for research \nstudies as well: illustrations in a paper helps the reader to better understand \nthe findings of the authors. There are already several possibilities for visualizing \ndata. But there always exist cases when the currently available diagram \ntypes are not useful enough. We also ran into such a situation, and created \ntwo new diagram types: Cumulative Characteristic Diagram and Quantile \nDifference Diagram for illustrating data sets of numeric types. \nThe Cumulative Characteristic Diagram is a curve, which is based on the \nnon-ascending order of the values. It makes it easy to read many characteristics \nof the input data, and it is suitable to find similarities and differences \nbetween several data sets quickly. \nQuantile Difference Diagram draws the differences of two ascending sets \nof data on the same quantiles. This diagram is suitable to illustrate in which \nsubset the data are higher, and it also reveals some important details, which \nwould remain hidden using statistic tests only. \nWe found them very useful both in explaining our actual results, and \ngaining ideas for further development directions. In this article we show \nthe usefulness of these diagrams illustrating the results of Contingency ChiSquared \ntests, Wilcoxon rank tests and variance tests.
Author: Csaba Faragó
Institution: University of Szeged
Language: English
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2567314018
