Generalized low rank models
From MaRDI portal
Abstract: Principal components analysis (PCA) is a well-known technique for approximating a tabular data set by a low rank matrix. Here, we extend the idea of PCA to handle arbitrary data sets consisting of numerical, Boolean, categorical, ordinal, and other data types. This framework encompasses many well known techniques in data analysis, such as nonnegative matrix factorization, matrix completion, sparse and robust PCA, -means, -SVD, and maximum margin matrix factorization. The method handles heterogeneous data sets, and leads to coherent schemes for compressing, denoising, and imputing missing entries across all data types simultaneously. It also admits a number of interesting interpretations of the low rank factors, which allow clustering of examples or of features. We propose several parallel algorithms for fitting generalized low rank models, and describe implementations and numerical results.
For large datasets, the task of processing data by different techniques proves difficult because of the large dataset dimension. One way to tackle this issue is the use of PCA by transforming the original variables to a smaller number of uncorrelated variables. This monograph is inspired by this paradigm, introducing a template method which embeds both the items (rows) and the features (columns) of the database into the same low-dimensional vector space, regardless of the data type. Many well-known techniques in data analysis are used (e.g., non-negative matrix factorization, sparse and robust PCA, \(k\)-means, etc.). Moreover, there are implementations of these techniques using Python, Julia and Spark tools.
Recommendations
Cited in
(54)- Leveraging mixed and incomplete outcomes via reduced-rank modeling
- Main effects and interactions in mixed and incomplete data frames
- ePCA: high dimensional exponential family PCA
- Flexible low-rank statistical modeling with missing data and side information
- Hyperparameter optimization for recommender systems through Bayesian optimization
- An objective penalty function method for biconvex programming
- Low-rank matrix denoising for count data using unbiased Kullback-Leibler risk estimation
- Acceptable set topic modeling
- Kurdyka-Łojasiewicz exponent via inf-projection
- Inertial alternating direction method of multipliers for non-convex non-smooth optimization
- Enhanced image approximation using shifted rank-1 reconstruction
- Dimensionality reduction for binary data through the projection of natural parameters
- Imputation and low-rank estimation with missing not at random data
- Bayesian mean-parameterized nonnegative binary matrix factorization
- Low-rank matrix approximation in the infinity norm
- Matrix optimization over low-rank spectral sets: stationary points and local and global minimizers
- Rank regularized estimation of approximate factor models
- Multiple correspondence analysis and the multilogit bilinear model
- A survey on deep matrix factorizations
- SONIC: social network analysis with influencers and communities
- Dynamic assortment personalization in high dimensions
- LowRankModels
- Pyglrm
- A nonmonotone alternating updating method for a class of matrix factorization problems
- Generalized SURE for optimal shrinkage of singular values in low-rank matrix denoising
- Adapting regularized low-rank models for parallel architectures
- Exponential-Family Embedding With Application to Cell Developmental Trajectories for Single-Cell RNA-Seq Data
- Simplex-Structured Matrix Factorization: Sparsity-Based Identifiability and Provably Correct Algorithms
- Why Are Big Data Matrices Approximately Low Rank?
- Stochastic gradients for large-scale tensor decomposition
- Mixed-Projection Conic Optimization: A New Paradigm for Modeling Rank Constraints
- A nonlinear matrix decomposition for mining the zeros of sparse data
- Real-Time Radiation Treatment Planning with Optimality Guarantees via Cluster and Bound Methods
- Structured latent factor analysis for large-scale data: identifiability, estimability, and their implications
- Non-convex matrix completion and related problems via strong duality
- Collective matrix completion
- Generalized canonical polyadic tensor decomposition
- On a Problem of Weighted Low-Rank Approximation of Matrices
- Understanding complex datasets. Data mining with matrix decompositions
- Signal Decomposition Using Masked Proximal Operators
- Embedding Learning
- Nonnegative decomposition of functional count data
- Representation in Machine Learning
- Exterior-point optimization for sparse and low-rank optimization
- Coherent set identification via direct low rank maximum likelihood estimation
- Generalized Principal Component Analysis: Projection of Saturated Model Parameters
- Parameter inference from a non-stationary unknown process
- Solving matrix nearness problems via Hamiltonian systems, matrix factorization, and optimization
- Frequency-aware tensor network decomposition for multi-dimensional image recovery
- Mixed Matrix Completion in Complex Survey Sampling under Heterogeneous Missingness
- Dual simplex volume maximization for simplex-structured matrix factorization
- Smooth optimization using global and local low-rank regularizers
- Latent structural categorical matrix completion with application to quasispecies analysis
- High-dimensional analysis of ridge regression for non-identically distributed data with a variance profile
This page was built for publication: Generalized low rank models
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2825734)
