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

Binary Classifier Calibration Using an Ensemble of Linear Trend Estimation.

Mahdi Pakdaman Naeini1, Gregory F Cooper2

  • 1Intelligent Systems Program, University of Pittsburgh.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining
|March 31, 2017
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 author

The use of logic for machine learning models in sepsis.

Intensive care medicine experimental·2026
Same author

A Decision-Theoretic Perspective on Fairness in Clinical Predictive Models.

Research square·2026
Same author

Causal modeling reveals cell-cell communication dynamics in the tumor microenvironment during anti-PD-1 therapy in breast cancer patients.

Briefings in bioinformatics·2026
Same author

An evaluation of a Bayesian method to track outbreaks of known and novel influenza-like illnesses.

Scientific reports·2026
Same author

Leveraging Expert Knowledge and Causal Structure Learning to Build Parsimonious Models of Acute Brain Dysfunction in the Pediatric Intensive Care Unit (PICU).

medRxiv : the preprint server for health sciences·2026
Same author

Reply to Eccleston and Moore.

Pain·2026
Same journal

AnchorDrug: A system for drug-induced gene expression prediction in new contexts through active learning.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2026
Same journal

Domain-Adaptive Continual Meta-Learning for Modeling Dynamical Systems: An Application in Environmental Ecosystems.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2025
Same journal

MedDiffusion: Boosting Health Risk Prediction via Diffusion-based Data Augmentation.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2024
Same journal

Automated Fusion of Multimodal Electronic Health Records for Better Medical Predictions.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2024
Same journal

FAME: Fragment-based Conditional Molecular Generation for Phenotypic Drug Discovery.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2022
Same journal

Harmonic Alignment.

Proceedings of the ... SIAM International Conference on Data Mining. SIAM International Conference on Data Mining·2021
See all related articles

A new method, ensemble of linear trend estimation (ELiTE), improves probabilistic model calibration for data mining. ELiTE offers superior performance over existing methods, enhancing calibration without sacrificing discrimination power.

Area of Science:

  • Data Mining
  • Machine Learning
  • Probabilistic Modeling

Background:

  • Accurate probabilistic models are essential for data mining tasks.
  • Existing calibration methods like histogram binning have limitations, including piecewise constant mappings and independence assumptions.

Purpose of the Study:

  • Introduce a novel non-parametric calibration method, ensemble of linear trend estimation (ELiTE).
  • Address the limitations of current calibration techniques.
  • Improve the accuracy of probability estimates from classifiers.

Main Methods:

  • Utilize the ℓ1 trend filtering signal approximation for mapping uncalibrated scores to probabilities.
  • Post-process outputs of binary classifiers.
  • Implement a computationally tractable O(N log N) algorithm.

Related Experiment Videos

Main Results:

  • ELiTE outperforms common binary-classifier calibration methods on real datasets.
  • ELiTE demonstrates statistically significant improvements over other methods.
  • The method enhances calibration power while preserving discrimination power.

Conclusions:

  • ELiTE is an effective and computationally efficient method for probabilistic model calibration.
  • The approach is broadly applicable to various binary classification models.
  • ELiTE offers a significant advancement over traditional calibration techniques.