Constructing and Visualizing High-Quality Classifier Decision Boundary Maps
| Autor: | Alexandru Telea, Roberto Hirata, Francisco Caio M. Rodrigues, Mateus Espadoto |
|---|---|
| Přispěvatelé: | Scientific Visualization and Computer Graphics |
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
Computer science
media_common.quotation_subject 02 engineering and technology Machine learning computer.software_genre ENCODE image-based visualization 0202 electrical engineering electronic engineering information engineering EIGENMAPS NONLINEAR DIMENSIONALITY REDUCTION VISÃO COMPUTACIONAL media_common dimensionality reduction Creative visualization PROJECTION lcsh:T58.5-58.64 business.industry lcsh:Information technology Dimensionality reduction 020207 software engineering Trustworthiness machine learning Decision boundary 020201 artificial intelligence & image processing Artificial intelligence business Classifier (UML) computer Information Systems |
| Zdroj: | Information, Vol 10, Iss 9, p 280 (2019) Information Volume 10 Issue 9 AHF-Information, 10(9):280. MDPI AG Repositório Institucional da USP (Biblioteca Digital da Produção Intelectual) Universidade de São Paulo (USP) instacron:USP |
| ISSN: | 2078-2489 |
| Popis: | Visualizing decision boundaries of machine learning classifiers can help in classifier design, testing and fine-tuning. Decision maps are visualization techniques that overcome the key sparsity-related limitation of scatterplots for this task. To increase the trustworthiness of decision map use, we perform an extensive evaluation considering the dimensionality-reduction (DR) projection techniques underlying decision map construction. We extend the visual accuracy of decision maps by proposing additional techniques to suppress errors caused by projection distortions. Additionally, we propose ways to estimate and visually encode the distance-to-decision-boundary in decision maps, thereby enriching the conveyed information. We demonstrate our improvements and the insights that decision maps convey on several real-world datasets. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|