In the second part of the „Subspace clustering lecture“ we discuss more advanced visual exploration techniques. Dimension glyphs and Clust Nails are examples for such advanced techniques. Some techniques rely on dimension reduction where the data of the subspace is typically reduced to 2D to be shown in a scatterplot. These dimension reduction techniques are discussed in Lecture 7. We show some successful applications of subspace clusterings and discuss their evaluation, e.g., the assessment whether subspace clustering provides valuable and reliable information. Evaluation often includes artificially generated data, where the clusters are known, and real-world data, e.g., for crime analysis or from healthcare.
Chapters:
00:00 - Visualization
10:50 - Applications
26:20 - Epidemiological Data
32:39 - Visual Quality
40:29 - Evaluation
48:36 - From Subspace Clusters to Concepts
53:26 - Summary, Limitations and References
