Summary of Revisiting Silhouette Aggregation, by John Pavlopoulos et al.
Revisiting Silhouette Aggregation
by John Pavlopoulos, Georgios Vardakas, Aristidis Likas
First submitted to arxiv on: 11 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper investigates the Silhouette coefficient, an internal clustering evaluation measure that scores each data point based on its clustering assignment. Typically, the scores are micro-averaged into a single value to assess the quality of the whole dataset’s clustering. However, the authors argue that this approach is sensitive to cluster imbalance and propose an alternative strategy: macro-averaging at the cluster level before averaging across clusters. The paper demonstrates that macro-Silhouette is more robust against imbalance using synthetic and real-world datasets from eight different domains. To improve the measure’s performance, the authors develop a per-cluster sampling method to address the issue of uniform sub-sampling. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research paper looks at a way to measure how well data points are grouped together in a cluster. The current method of measuring this is sensitive to differences in the number of points in each group, which can affect the results. The authors suggest an alternative approach that takes into account the size of each group and find it to be more reliable. They also propose a new way to randomly select points from each group to improve the accuracy of the measurements. |
Keywords
* Artificial intelligence * Clustering
