A Framework for Feature Selection in Clustering
Daniela Witten, Robert John Tibshirani · Journal of the American Statistical Association · 2010
We consider the problem of clustering observations using a potentially large set of features. One might expect that the true underlying clusters present in the data differ only with respect to a small fraction of the features, and will be missed if one clusters the observations using the full set of features. We propose a novel framework for sparse clustering, in which one clusters the observations using an adaptively chosen subset of the features. The method uses a lasso-type penalty to select the features. We use this framework to develop simple methods for sparse K-means and sparse hierarchical clustering. A single criterion governs both the selection of the features and the resulting clusters. These approaches are demonstrated on simulated data and on genomic data sets.
Cited by 696 works. doi:10.1198/jasa.2010.tm09415
Authors
References
- Maximum Likelihood from Incomplete Data Via the EM Algorithm
- Convex Optimization
- Cluster analysis and display of genome-wide expression patterns
- Some methods for classification and analysis of multivariate observations
- Molecular portraits of human breast tumours
- Learning the parts of objects by non-negative matrix factorization
- Finding groups in data: an introduction to cluster analysis
- A haplotype map of the human genome
- Principal components analysis corrects for stratification in genome-wide association studies
- Objective Criteria for the Evaluation of Clustering Methods