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 flexible coefficient smooth transition time series model.

Marcelo C Medeiros1, Alvaro Veiga

  • 1Department of Economics, Pontifical Catholic University of Rio de Janeiro, Rio de Janeiro, RJ 22451-900 Brazil.

IEEE Transactions on Neural Networks
|March 1, 2005
PubMed
Summary

This study introduces a flexible smooth transition autoregressive (STAR) model using neural networks. This advanced model unifies various nonlinear time series models, demonstrating effectiveness in simulations.

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

Modeling the evolution of deaths from infectious diseases with functional data models: The case of COVID-19 in Brazil.

Statistics in medicine·2023
Same author

Short-term Covid-19 forecast for latecomers.

International journal of forecasting·2021
Same author

A Case of Paraneoplastic Cerebellar Degeneration that Preceded the Diagnosis of Classical Hodgkin's Lymphoma by 16 Months.

The American journal of case reports·2020
Same author

Revisiting hydrocarbons source appraisal in sediments exposed to multiple inputs.

Marine pollution bulletin·2013
Same author

Validation of Ucides cordatus as a bioindicator of oil contamination and bioavailability in mangroves by evaluating sediment and crab PAH records.

Environment international·2006

Area of Science:

  • Statistics
  • Econometrics
  • Machine Learning

Background:

  • Nonlinear time series analysis is crucial for modeling complex economic and financial data.
  • Existing models like SETAR, AR-NN, and logistic STAR have limitations in flexibility.
  • The need for models that can capture intricate, time-varying relationships is evident.

Purpose of the Study:

  • To propose a novel flexible smooth transition autoregressive (STAR) model incorporating multiple regimes and transition variables.
  • To demonstrate that this formulation nests several existing nonlinear time series models.
  • To develop a statistically sound model-building procedure for this new class of models.

Main Methods:

  • The proposed model uses a single hidden layer feedforward neural network to generate time-varying coefficients.

Related Experiment Videos

  • This approach allows for a flexible, nonparametric representation of the transition dynamics.
  • Model building is based on statistical inference, validated through Monte Carlo simulations.

    • Main Results:

    • The proposed flexible STAR model successfully nests SETAR, AR-NN, and logistic STAR models.
    • The model-building procedure demonstrates good performance in small and medium-sized samples.
    • The approach is comparable to functional coefficient autoregressive (FAR) and single-index coefficient regression models.

    Conclusions:

    • The flexible STAR model offers a powerful and unified framework for nonlinear time series analysis.
    • The developed statistical inference procedure is effective for practical model implementation.
    • This research advances the capabilities of time series modeling in econometrics and machine learning.