Abstract: The k-means algorithm with cosine similarity, also known as the spherical k-means algorithm, is a popular method for clustering document collections. However, spherical k-means can often yield qualitatively poor results, especially for small clusters, say 25-30 documents per cluster, where it tends to get stuck at a local maximum far away from the optimal. In this paper, we present the first-variation principle that refines a given clustering by incrementally moving data points between clusters, thus achieving a higher objective function value. Combining first-variation with spherical k-means yields a powerful ping-pong strategy that often qualitatively improves k-means clustering. We present several experimental results to show that our proposed method works well in clustering high-dimensional and sparse text data.
- Data Clustering
- Refining Clusters in High-dimensional Text Data (pdf)
I. Dhillon, Y. Guan, J. Kogan.
In SIAM International Conference on Data Mining (SDM), April 2002.