Abstract | ||
---|---|---|
textabstractMost multi-class classifiers make their prediction for a test sample by scoring theclasses and selecting the one with the highest score. Analyzing these predictionscores is useful to understand the classifier behavior and to assess its reliability. Wepresent an interactive visualization that facilitates per-class analysis of these scores.Our system, called Classilist, enables relating these scores to the classificationcorrectness and to the underlying samples and their features. We illustrate how suchanalysis reveals varying behavior of different classifiers. Classilist is available foruse online, along with source code, video tutorials, and plugins for R, RapidMiner,and KNIME at https://katehara.github.io/classilist-site/. |
Year | Venue | Field |
---|---|---|
2017 | arXiv: Machine Learning | Data mining,Source code,Computer science,Correctness,Interactive visualization,Artificial intelligence,Plug-in,Classifier (linguistics),Machine learning |
DocType | Volume | Citations |
Journal | abs/1711.06795 | 1 |
PageRank | References | Authors |
0.35 | 2 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Medha Katehara | 1 | 1 | 0.35 |
Emma Beauxis-Aussalet | 2 | 20 | 4.75 |
Bilal Alsallakh | 3 | 196 | 11.71 |