Comparative Study of Inference Methods for Interpolative Decomposition

@article{Lu2022ComparativeSO,
  title={Comparative Study of Inference Methods for Interpolative Decomposition},
  author={Jun Lu},
  journal={ArXiv},
  year={2022},
  volume={abs/2206.14542}
}
  • Jun Lu
  • Published 29 June 2022
  • Computer Science
  • ArXiv
We propose a probabilistic model with automatic relevance determination (ARD) for computing interpolative decomposition (ID), that is commonly used for low-rank approximation, feature selection, and extracting hidden patterns in data, where the matrix factors are latent variables associated with each data dimension. The constraint on the magnitude of the factored component is addressed using prior densities with support on the designated subspace. Bayesian inference procedure based on Gibbs… 

Figures and Tables from this paper

Robust Bayesian Nonnegative Matrix Factorization with Implicit Regularizers

A probabilistic model with implicit norm regularization for learning nonnegative matrix factorization (NMF) that is commonly used for predicting missing values and hidden patterns in the data is introduced and proposed Bayesian NMF GL 22 and GL 22, ∞ models lead to robust predictions for different data values and avoid over-prediction compared with competitive BayesianNMF approaches.

Feature Selection via the Intervened Interpolative Decomposition and its Application in Diversifying Quantitative Strategies

In this paper, we propose a probabilistic model for computing an interpolative decomposition (ID) in which each column of the observed matrix has its own priority or importance, so that the end

References

SHOWING 1-10 OF 28 REFERENCES

Bayesian Low-Rank Interpolative Decomposition for Complex Datasets

  • Jun Lu
  • Computer Science
    Studies in Engineering and Technology
  • 2022
In this paper, we introduce a probabilistic model for learning interpolative decomposition (ID), which is commonly used for feature selection, low-rank approximation, and identifying hidden patterns

Probabilistic interpolative decomposition

  • I. AriA. CemgilL. Akarun
  • Computer Science
    2012 IEEE International Workshop on Machine Learning for Signal Processing
  • 2012
This work proposes a novel probabilistic method for ID where it is expressed as a statistical model within a Bayesian framework and enables the construction of problem-specific interpolative decompositions.

Flexible and Hierarchical Prior for Bayesian Nonnegative Matrix Factorization

A probabilistic model for learning nonnegative matrix factorization (NMF) that is commonly used for predicting missing values and hidden patterns in the data, in which the matrix factors are latent variables as-sociated with each data dimension is introduced.

A survey on Bayesian inference for Gaussian mixture model

  • Jun Lu
  • Computer Science
    ArXiv
  • 2021
This survey is primarily a summary of purpose, significance of important background and techniques for Gaussian mixture model, e.g., Dirichlet prior, Chinese restaurant process, and most importantly the origin and complexity of the methods which shed light on their modern applications.

Principal Component Analysis for Large Scale Problems with Lots of Missing Values

This work proposes an algorithm based on speeding up a simple principal subspace rule, and extends it to use regularization and variational Bayesian learning, and finds that VB-PCA method provides more accurate predictions for new data than traditional PCA or regularized PCA.

Prior and Likelihood Choices for Bayesian Matrix Factorisation on Small Datasets

In this paper, several insights are offered into the trade-offs between prior and likelihood choices for Bayesian matrix factorisation on small datasets - such as that Poisson models give poor predictions, and that nonnegative models are more constrained than real-valued ones.

Variational Bayesian Approach to Movie Rating Prediction

A Bayesian approach is proposed to alleviate overfitting in SVD, where priors are introduced and all parameters are integrated out using variational inference, and it is shown that this gives significantly improved results over vanilla SVD.

Scalable Recommendation with Hierarchical Poisson Factorization

H hierarchical Poisson matrix factorization is developed, a novel method for providing users with high quality recommendations based on implicit feedback, such as views, clicks, or purchases, and it is shown that it more accurately captures the long-tailed user activity found in most consumption data.

Bayesian Nonparametric Poisson Factorization for Recommendation Systems

A Bayesian nonparametric Poisson factorization model for recommendation systems that eases the computational burden of searching for the number of latent components and gives better predictive performance than its parametric counterpart.

Numerical Matrix Decomposition and its Modern Applications: A Rigorous First Course

  • Jun Lu
  • Computer Science
    ArXiv
  • 2021
The sole aim of this survey is to give a self-contained introduction to concepts and mathematical tools in numerical linear algebra and matrix analysis in order to seamlessly introduce matrix decomposition techniques and their applications in subsequent sections.