Identifying Sparse Low-Dimensional Structures in Markov Chains: A Nonnegative Matrix Factorization Approach.
Mahsa GhasemiAbolfazl HashemiHaris VikaloUfuk TopcuPublished in: ACC (2020)
Keyphrases
- markov chain
- nonnegative matrix factorization
- low dimensional
- high dimensional
- sparsity constraints
- principal component analysis
- negative matrix factorization
- matrix factorization
- transition probabilities
- state space
- least squares
- data representation
- dimensionality reduction
- probabilistic automata
- random walk
- original data
- markov processes
- spectral clustering
- transition matrix
- sparse representation
- data points
- high dimensional data
- feature space
- clustering method
- document clustering
- objective function
- image processing
- dictionary learning
- machine learning
- relational databases
- pattern recognition