AI-ordlista
Den kompletta ordlistan över AI
Singular Value Decomposition (SVD)
Fundamental matrix factorization method decomposing a matrix into three orthogonal matrices to reduce dimensionality.
Non-Negative Matrix Factorization (NMF)
A decomposition technique where all elements of the factor matrices are constrained to be non-negative, ideal for interpretability.
Probabilistic Matrix Factorization (PMF)
Bayesian approach to factorization modeling data with probabilistic distributions to handle uncertainty.
Tensor Matrix Factorization
Extension of matrix factorization to multidimensional tensors for analyzing data with more than two dimensions.
Matrix Factorization with Regularization
Incorporation of regularization terms (L1, L2) to prevent overfitting and improve model generalization.
Online Matrix Factorization
Adaptive algorithms updating matrix factors in real-time with the arrival of new data.
Distributed Matrix Factorization
Parallel approaches for factorizing massive matrices on distributed systems such as Spark or Hadoop.
Matrix Factorization for Missing Data
Specialized techniques for matrix completion with numerous missing values, typical in recommendation systems.
Robust Matrix Factorization
Methods resistant to outliers and noise in the initial data for a more stable decomposition.
Matrix Factorization with Constraints
Incorporation of specific constraints (spatial, temporal, semantic) to guide factorization according to the application domain.
Hierarchical Matrix Factorization
Multi-level approaches capturing hierarchical structures in data for a richer representation.
Matrix Factorization for Time Series
Techniques adapted to decompose temporal data into trends, seasonalities, and latent components.