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

Nonlinear time series analysis by neural networks: a case study

H Saxén

    International Journal of Neural Systems
    |May 1, 1996
    PubMed
    Summary

    This study explores neural networks for nonlinear time-series analysis. Optimal network sizing prevents overfitting and improves generalization for univariate systems.

    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

    Replacing risk-based early-onset-disease prevention with intrapartum group B streptococcus PCR testing.

    The journal of maternal-fetal & neonatal medicine : the official journal of the European Association of Perinatal Medicine, the Federation of Asia and Oceania Perinatal Societies, the International Society of Perinatal Obstetricians·2016
    Same author

    Infant botulism with prolonged faecal excretion of botulinum neurotoxin and Clostridium botulinum for 7 months.

    Epidemiology and infection·2013
    Same author

    Effect of adenoidectomy on respiratory function: a randomised prospective study.

    Archives of disease in childhood·2009
    Same author

    Nosocomial bloodstream infections in children: an 8-year experience at a tertiary-care hospital in Finland.

    Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases·2008
    Same author

    A randomized, controlled trial of tonsillectomy in periodic fever, aphthous stomatitis, pharyngitis, and adenitis syndrome.

    The Journal of pediatrics·2007
    Same author

    [New antibacterial agents].

    Duodecim; laaketieteellinen aikakauskirja·2002

    Area of Science:

    • Computational neuroscience
    • Nonlinear dynamics
    • Machine learning

    Background:

    • Time-series analysis is crucial for understanding dynamic systems.
    • Neural networks offer a powerful tool for modeling complex, nonlinear behaviors.
    • Determining optimal neural network architecture is essential for accurate predictions.

    Discussion:

    • This research investigates feedforward neural networks for univariate nonlinear time-series analysis.
    • Network size selection is based on performance metrics from training and testing datasets.
    • Phase portrait analysis validates the models' ability to capture system dynamics.

    Key Insights:

    • Appropriate neural network sizing effectively avoids over-parameterization issues.
    • Larger networks exhibit overfitting, leading to diminished generalization capabilities.
    • Incomplete training in oversized networks results in approximations distinct from optimal models.

    Outlook:

    • Further research can explore recurrent neural networks for more complex time-series.
    • Investigating different regularization techniques can enhance generalization further.
    • Applying these findings to real-world nonlinear systems could yield significant insights.

    Related Experiment Videos