We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Interactive Dimensionality Reduction for Comparative Analysis

Abstract: Finding the similarities and differences between groups of datasets is a fundamental analysis task. For high-dimensional data, dimensionality reduction (DR) methods are often used to find the characteristics of each group. However, existing DR methods provide limited capability and flexibility for such comparative analysis as each method is designed only for a narrow analysis target, such as identifying factors that most differentiate groups. This paper presents an interactive DR framework where we integrate our new DR method, called ULCA (unified linear comparative analysis), with an interactive visual interface. ULCA unifies two DR schemes, discriminant analysis and contrastive learning, to support various comparative analysis tasks. To provide flexibility for comparative analysis, we develop an optimization algorithm that enables analysts to interactively refine ULCA results. Additionally, the interactive visualization interface facilitates interpretation and refinement of the ULCA results. We evaluate ULCA and the optimization algorithm to show their efficiency as well as present multiple case studies using real-world datasets to demonstrate the usefulness of this framework.
Comments: This is the author's version of the article that has been published in IEEE Transactions on Visualization and Computer Graphics. The final version of this record is available at: 10.1109/TVCG.2021.3114807
Subjects: Machine Learning (cs.LG); Human-Computer Interaction (cs.HC); Machine Learning (stat.ML)
DOI: 10.1109/TVCG.2021.3114807
Cite as: arXiv:2106.15481 [cs.LG]
  (or arXiv:2106.15481v3 [cs.LG] for this version)

Submission history

From: Takanori Fujiwara [view email]
[v1] Tue, 29 Jun 2021 15:05:36 GMT (6380kb,D)
[v2] Sun, 8 Aug 2021 18:29:36 GMT (6828kb,D)
[v3] Wed, 27 Oct 2021 16:22:41 GMT (6828kb,D)

Link back to: arXiv, form interface, contact.