Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A variational EM approach to predictive uncertainty.

Markus Harva1

  • 1Laboratory of Computer and Information Science, Helsinki University of Technology, Espoo, Finland. markus.harva@tkk.fi

Neural Networks : the Official Journal of the International Neural Network Society
|May 26, 2007
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same journal

Anchor-based disentanglement framework for incremental multi-view clustering.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Complex-valued amplitude-phase interference modeling for adversarially robust classification.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

TraNce: Type-aware hypergraph neural network with biological mediators for drug repositioning.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Decentralized ADMM for factorization-based Low-rank matrix estimation.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Memristive neuromorphic circuit design inspired by the neural mechanisms of conditioned fear.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Q-learning based asynchronous Boolean control networks stabilization with data loss.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

Predicting uncertainty in regression requires modeling the full conditional probability density. This method conditions the noise process scale on explanatory variables, using variational EM to prevent overfitting and improve predictive distributions, as validated in a competition.

Area of Science:

  • Machine Learning
  • Statistical Modeling
  • Environmental Science

Background:

  • Traditional regression often relies on conditional averages, which are insufficient for complex dependencies.
  • Modeling the full conditional probability density is crucial but prone to overfitting.

Purpose of the Study:

  • To develop a robust method for predicting uncertainty in regression by modeling the conditional probability density.
  • To address overfitting issues in flexible density estimation models.

Main Methods:

  • Conditioning the scale parameter of the noise process on explanatory variables.
  • Utilizing variational Expectation-Maximization (EM) for model learning.
  • Developing a model where unpredictability in scale enhances predictive distribution robustness.

Related Experiment Videos

Main Results:

  • The proposed method effectively models conditional probability density and predicts uncertainty.
  • Variational EM successfully mitigates overfitting problems.
  • Experimental validation on synthetic and real-world environmental data demonstrated viability.

Conclusions:

  • The developed approach provides a robust solution for uncertainty prediction in regression tasks.
  • The method's effectiveness was confirmed by winning the 'Predictive uncertainty in environmental modelling' competition.