Model Comparisons

Tests for nested models and information criteria

After you estimate a set of candidate models, you can compute information criteria to compare the model fits by passing the resulting set of loglikelihoods to the aicbic function. The Econometric Modeler app enables you to interactively use compare models using information criteria. Alternatively, you can conduct statistical hypothesis tests to compare the fitted models, for example, conduct a likelihood ratio test to compare nested models by using the lratiotest function.

Apps

Econometric Modeler

Analyze and model econometric time series

Functions

expand all

Information Criteria

aicbic Information criteria

Statistical Tests

`lmtest`	Lagrange multiplier test of model specification
`lratiotest`	Likelihood ratio test of model specification
`waldtest`	Wald test of model specification

Convert Between Models

`arma2ar`	Convert ARMA model to AR model
`arma2ma`	Convert ARMA model to MA model
`var2vec`	Convert VAR model to VEC model
`vec2var`	Convert VEC model to VAR model

Topics

Information Criteria

Information Criteria for Model Selection
Compare model fits using information criteria.
Choose ARMA Lags Using BIC
Select ARMA model using information criteria.
Compare Conditional Variance Model Fit Statistics Using Econometric Modeler App
Interactively specify and fit GARCH, EGARCH, and GJR models to data. Then, determine the model that fits to the data the best by comparing fit statistics.
Compare Conditional Variance Models Using Information Criteria
Compare the fits of several conditional variance models using AIC and BIC.
Select Regression Model with ARIMA Errors
Learn how to select an appropriate regression model with ARIMA errors.
Rolling-Window Analysis of Time-Series Models
Estimate explicitly and implicitly defined state-space models using a rolling window.
Choose State-Space Model Specification Using Backtesting
Choose the state-space model specification with the best predictive performance using a rolling window.

Statistical Tests

Model Comparison Tests
Learn the mechanics behind the likelihood ratio, Lagrange multiplier, and Wald model-comparison tests.
Compare GARCH Models Using Likelihood Ratio Test
Conduct a likelihood ratio test to choose the number of lags in a GARCH model.
Likelihood Ratio Test for Conditional Variance Models
Fit two competing, conditional variance models to data, and then compare their fits using a likelihood ratio test.
Conduct Lagrange Multiplier Test
Compare the fit of a restricted model against an unrestricted model by testing whether the gradient of the loglikelihood function of the unrestricted model, evaluated at the restricted maximum likelihood estimates (MLEs), is significantly different from zero.
Conduct Wald Test
Compare the fit of a restricted model against an unrestricted model by testing whether the restriction function, evaluated at the unrestricted maximum likelihood estimates (MLEs), is significantly different from zero.
Classical Model Misspecification Tests
This example shows the use of the likelihood ratio, Wald, and Lagrange multiplier tests.

Convert Between Models

Convert VARMA Model to VAR Model
Create a VARMA model, and then convert it to a pure VAR model.

Related Information

Featured Examples

Time Series Regression V: Predictor Selection

Select a parsimonious set of predictors with high statistical significance for multiple linear regression models. It is the fifth in a series of examples on time series regression, following the presentation in previous examples.

Open Live Script