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 measure of model reliability.

L R Williams, R W Leggett

    Health Physics
    |January 1, 1984
    PubMed
    Summary
    This summary is machine-generated.

    This study introduces a new method to assess predictive model accuracy within a factor of k. The developed reliability indices, kg and ks, offer a statistically sound way to evaluate model performance using observed data and predictions.

    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

    Occupational Intakes of Radionuclides: Part 5.

    Annals of the ICRP·2022
    Same author

    A photo-elicitation exploration of UK mothers' experiences of extended breastfeeding.

    Appetite·2021
    Same author

    ICRP Publication 141: Occupational Intakes of Radionuclides: Part 4.

    Annals of the ICRP·2019
    Same author

    A biokinetic model for trivalent or hexavalent chromium in adult humans.

    Journal of radiological protection : official journal of the Society for Radiological Protection·2019
    Same author

    Action levels for airborne uranium in the workplace: chemical and radiological assessments.

    Journal of radiological protection : official journal of the Society for Radiological Protection·2018
    Same author

    ICRP Publication 137: Occupational Intakes of Radionuclides: Part 3.

    Annals of the ICRP·2018
    Same journal

    Evaluation of Lung Doses from Radon and Thoron Exposure by Using LR-115 Detectors.

    Health physics·2026
    Same journal

    Radon and Radium Levels in Soil and Associated Indoor and Outdoor Radon Radiation Hazards in the Raparin Area, Sulaymaniyah, Iraqi Kurdistan.

    Health physics·2026
    Same journal

    Ionizing Radiation Hazard for Laser Material Processing Applications Using an Ultrashort Pulse Laser.

    Health physics·2026
    Same journal

    Assessment of Health Risks of Adults and Children Due to Consumption of Uranium in Groundwater from Chengalpattu District, Tamil Nadu, India.

    Health physics·2026
    Same journal

    Radiation Protection Abstracts, Volume 46, Number 1.

    Health physics·2026
    Same journal

    Specialized Radiological Assets for Navigable Two-dimensional and Three-dimensional Virtual and Augmented Reality.

    Health physics·2026
    See all related articles

    Area of Science:

    • Statistics
    • Predictive Modeling
    • Data Analysis

    Background:

    • Evaluating the accuracy of predictive models is crucial for reliable scientific insights.
    • Existing methods may lack precise statistical interpretation for accuracy within a factor 'k'.
    • A need exists for robust reliability indices applicable across various models.

    Purpose of the Study:

    • To develop a statistically meaningful method for interpreting predictive model accuracy within a factor of k.
    • To introduce new reliability indices (kg and ks) for model evaluation.
    • To explore the statistical properties and applications of these indices.

    Main Methods:

    • Defined a geometrically intuitive reliability index (kg) based on the ratios of observed values to predicted values (yi/xi).

    Related Experiment Videos

  • Defined a statistically rigorous reliability index (ks) using the natural logarithm of these ratios.
  • Investigated the relationship between kg and ks and the underlying model parameter exp[(V1 + V2)1/2].

    • Main Results:

    • Introduced kg and ks as interchangeable reliability indices for accurate models.
    • ks estimates a parameter reflecting observational variance (V1) and model uncertainty (V2).
    • The probability distribution of ks can be characterized under specific conditions (lognormal, independent, homoscedastic distributions).

    Conclusions:

    • The developed method provides a precise and statistically sound interpretation of model accuracy within a factor of k.
    • The reliability indices kg and ks offer a valuable tool for assessing predictive model performance.
    • The method is particularly relevant for radiation risk analyses where data often meets the required distributional assumptions.