Non-negative Matrix Factorization
GPTKB entity
Statements (48)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:algorithm
|
| gptkbp:abbreviation |
gptkb:NMF
|
| gptkbp:application |
bioinformatics
image processing recommender systems text mining |
| gptkbp:category |
gptkb:matrix_decomposition
unsupervised learning |
| gptkbp:field |
gptkb:machine_learning
linear algebra data mining |
| gptkbp:input_constraint |
input matrix must have non-negative elements
|
| gptkbp:introduced |
1999
Daniel D. Lee H. Sebastian Seung |
| gptkbp:limitation |
local minima
scalability for large datasets sensitivity to initialization |
| gptkbp:objective_function |
minimize Frobenius norm
minimize Kullback-Leibler divergence minimize reconstruction error |
| gptkbp:optimizedFor |
alternating least squares
multiplicative update rules |
| gptkbp:output |
approximate factorization
two non-negative matrices |
| gptkbp:prohibits |
non-negativity
|
| gptkbp:property |
interpretable factors
non-uniqueness of solution parts-based representation |
| gptkbp:relatedConcept |
gptkb:Independent_Component_Analysis
gptkb:Factor_Analysis Convex NMF Dictionary Learning Non-negative Tensor Factorization Sparse NMF |
| gptkbp:relatedTo |
gptkb:Singular_Value_Decomposition
gptkb:Latent_Semantic_Analysis Principal Component Analysis |
| gptkbp:software |
gptkb:TensorFlow
gptkb:MATLAB gptkb:scikit-learn R |
| gptkbp:usedFor |
dimensionality reduction
feature extraction topic modeling |
| gptkbp:bfsParent |
gptkb:NMF
|
| gptkbp:bfsLayer |
8
|
| https://www.w3.org/2000/01/rdf-schema#label |
Non-negative Matrix Factorization
|