Permvizor:Visual Analysis Of Multivariate Permutations

Permutation is one of the most common and basic topics in combinatorics.Many problems arise in the fields of science,engineering and business.It is also found in mathematics,combinatorial mathematics,computer science,operations research,computational molecular biology,natural language processing,and even color schemes.However,the permutation problem appears to be a NP-complete problem,the number of permutations increases dramatically along with the number of items in the permutation.It takes a lot of time for users to manually test and measure all of the permutations,making it challenging for users to evaluate potential solutions and identify interesting insights.In order to solve the above shortcomings,the visualization analysis technology is used to solve the problem of multivariate permutations.The visualization technology is used to explore the process of searching the best sequence from the array,compare the differences between the sequences and find some patterns from the array,so as to help users make decisions.In this paper,we propose PermVizor,a novel and scalable visualization system that aims assisting users exploring the arrangement,distribution,and comparison of permutations.Eight necessary and comprehensive analysis of requirements are presented for the visualization of permutations.The main work and research results of the thesis are summarized as follows:(1)Data preprocessing and model construction: According to the characteristics of multivariate fully arranged data,a new similar distance calculation method is proposed.The multivariate data studied in this paper is reduced to the two-dimensional plane by MDS algorithm.Then according to the spatial distance of the scatter in the two-dimensional plane,clustering according to the DBSCAN algorithm.(2)Design visual components according to user requirements: The MDS view provides an overview of all sequences in the permutation,based on the extracted feature vectors distributed and clustered in the two-dimensional view.Allows users to explore the overall distribution of each permutation using primitives;the PCP view is designed to assist the user in viewing the statistics of the selected permutation and to examine the detailed arrangement of the items and the permutations of the permutations through pixel-based and block-based PermView.Multivariate information;the PermView view provides a variety of exploration modes to assist in analyzing the differences between sequence schemes.(3)System construction and case study: This work finally completed a complete Web system PermVizor,to meet the user's eight needs.The usefulness of the proposed visualization technique was further evaluated through case studies and qualitative studies of classical data sets.And the study shows that PermVizor can help users explore the alignment and confirm their discovery and decision making in the expected alignment.