Multidimensional Data Visualization Based on the Shortest Unclosed Path Search

The paper considers methods of multidimensional data visualization based on the search for the shortest unclosed path between the objects of the study sample and its mapping to a two-dimensional plane as an unclosed graph (chain), a columnar chart of the objects distribution along the found path or projection onto the path. The shortest unclosed path computing is executed with distance matrix between objects, which allows to use the method for both multidimensional data and data represented only by paired comparisons of objects functions. In this work an algorithm for greedy search of a quasi-shortest unclosed path is implements and also its modifications are proposed. The algorithms are tested on model and real-world data and also data obtained during the study of the problem of detecting falls using Microsoft Kinect V2.