Element-Wise Alternating Least Squares Algorithm for Nonnegative Matrix Factorization on One-Hot Encoded Data.
Zhuo WuTsuyoshi MigitaNorikazu TakahashiPublished in: ICONIP (5) (2020)
Keyphrases
- nonnegative matrix factorization
- alternating least squares
- spectral clustering
- input data
- data sets
- original data
- data analysis
- low rank approximation
- negative matrix factorization
- k means
- data points
- learning algorithm
- pairwise
- high dimensional data
- raw data
- convergence rate
- matrix factorization
- data mining
- objective function
- machine learning
- missing data
- probabilistic model
- optical flow