Spectral clustering is an alternative clustering technique that is rooted in graph theory. It works by constructing an adjacency matrix that connects similar points within a neighborhood defined by something like an epsilon radius or k-neighbors. It then projects the data into a lower dimensional space and applies a traditional clustering technique like K-Means on the projected data. Spectral clustering is particularly well-suited for finding non-convex clusters, such as the case where one cluster is embedded around a ring of data points that form a separate cluster.
What is Spectral Clustering?
- Categories: Unsupervised Learning
- Updated: March 26, 2023
Help us improve this post by suggesting in comments below: - modifications to the text, and infographics
- video resources that offer clear explanations for this question
- code snippets and case studies relevant to this concept
- online blogs, and research publications that are a “must read” on this topic
- video resources that offer clear explanations for this question
- code snippets and case studies relevant to this concept
- online blogs, and research publications that are a “must read” on this topic
Partner Ad
- Other Questions in Unsupervised Learning
Find out all the ways
that you can