Model Selection
Specification testing and model assessment
To begin selecting models for time series data, conduct hypothesis tests for stationarity, autocorrelation, and heteroscedasticity. After estimating the models, compare the fits using, for example, information criteria or a likelihood ratio test. You can also assess whether the models violate any assumptions by analyzing the residuals. For a multiple linear regression model, you can assess whether there is a structural change in the model, or address heteroscedasticity when estimating the regression coefficients.
Model Selection Basics
- Select ARIMA Model for Time Series Using Box-Jenkins Methodology
- Classical Model Misspecification Tests
- Time Series Regression I: Linear Models
- Time Series Regression V: Predictor Selection
- Time Series Regression IX: Lag Order Selection
- Likelihood Ratio Test for Conditional Variance Models
- Cointegration and Error Correction Analysis
Categories
- Specification Testing
Identify the parametric form of a model
- Model Comparisons
Tests for nested models and information criteria
- Residual Diagnostics
Evaluate model fit and performance
- Nonspherical Models
Model or correct effects of heteroscedasticity and correlation