Type of Publication: Article in Collected Edition

Visual Comparison of Multi-label Classification Results

Author(s):

Krause, Cedric; Agarwal, Shivam; Ghoniem, Mohammad; Beck, Fabian

Title of Anthology:

Vision, Modeling, and Visualization

Publication Date:

2021

Digital Object Identifier (DOI):

doi:10.2312/vmv.20211367

Fulltext:

Visual Comparison of Multi-label Classification Results (1.66 MB)

Citation:

Download BibTeX

Abstract

In multi-label classification, we do not only want to analyze individual data items but also the relationships between the assigned labels. Employing different sources and algorithms, the label assignments differ. We need to understand these differences to identify shared and conflicting assignments. We propose a visualization technique that addresses these challenges. In graphs, we present the labels for any classification result as nodes and the pairwise overlaps of labels as links between them. These graphs are juxtaposed for the different results and can be diffed graphically. Clustering techniques are used to further analyze similarities between labels or classification results, respectively. We demonstrate our prototype in two application examples from the machine learning domain.