Big Data Subspace Clustering

Project description/goals

Large-scale subspace clustering via k-factorization

KDD’21 (The 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining). 2021.

Importance/impact, challenges/pain points

1.Clustering 100k data points in 1 minute

2.Online clustering

Solution description

Group-sparse matrix factorization for clustering

Key contribution/commercial implication

1.Linear time and space complexity

2.High-clustering accuracy on large-scale datasets

3.Handle outliers and missing values

Team/contributors

Jicong Fan

Numerical results

Projects