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

Artificial neural networks for infant mortality modelling.

Ronaldo C Gismondi1, Renan Moritz Varnier R Almeida, Antonio Fernando C Infantosi

  • 1Medical Sciences College/State University of Rio de Janeiro, Rio de Janeiro, Brazil. renan@peb.ufrj.br

Computer Methods and Programs in Biomedicine
|September 3, 2002
PubMed
Summary

This study introduces a straightforward method for understanding artificial neural network (ANN) variable importance in epidemiological modeling. ANNs effectively predicted infant mortality using key social and economic factors, outperforming linear models.

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

Validation in principal components analysis applied to EEG data.

Computational and mathematical methods in medicine·2014
Same author

Assessing the combined performance of texture and morphological parameters in distinguishing breast tumors in ultrasound images.

Medical physics·2012
Same author

Peptides Derived from Mycobacterium leprae ML1601c Discriminate between Leprosy Patients and Healthy Endemic Controls.

Journal of tropical medicine·2012
Same author

Multiple Coherence vs Multiple Component Synchrony Measure for somatosensory evoked response detection.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2010
Same author

A non-linear morphometric feature selection approach for breast tumor contour from ultrasonic images.

Computers in biology and medicine·2010
Same author

Assessing the performance of morphological parameters in distinguishing breast tumors on ultrasound images.

Medical engineering & physics·2009

Area of Science:

  • Epidemiology
  • Data Science
  • Public Health

Background:

  • Artificial neural networks (ANNs) offer powerful predictive capabilities in complex modeling.
  • Assessing the relative importance of input variables in ANNs is crucial for interpretability and application.
  • Epidemiological modeling often requires understanding socio-economic and environmental determinants of health outcomes.

Purpose of the Study:

  • To develop and evaluate a user-friendly methodology for determining input variable importance in ANNs for epidemiological modeling.
  • To compare the predictive performance of ANNs against traditional multiple linear regression models.
  • To identify key socio-economic and environmental factors influencing infant mortality rates.

Main Methods:

  • Applied artificial neural networks (ANNs) to model infant mortality rates in 59 Brazilian municipalities.

Related Experiment Videos

  • Utilized factor analysis (FA) to select significant input variables from an initial set of 43.
  • Employed bootstrap replications to validate ANN model performance and assess variable importance.
  • Compared ANN models with multiple linear regression models (LRMs).

    • Main Results:

    • ANN models demonstrated superior performance, achieving R(2) values of 0.74-0.80, compared to LRMs (R(2)=0.4-0.5).
    • Bootstrap validation confirmed the higher accuracy and lower mean square error of ANN models.
    • Key predictors for infant mortality identified by the best ANN model included literacy, agricultural/livestock jobs, commercial establishments, and telephones.

    Conclusions:

    • The proposed methodology provides an interpretable approach to assessing ANN variable importance in public health.
    • ANN modeling shows significant potential for predicting, planning, and evaluating public health interventions.
    • Socio-economic factors play a critical role in determining infant mortality rates.