TD_BREUSCH_GODFREY Function | Teradata Vantage - TD_BREUSCH_GODFREY - Teradata Vantage

Database Unbounded Array Framework Time Series Functions

Deployment
VantageCloud
VantageCore
Edition
Enterprise
IntelliFlex
VMware
Product
Teradata Vantage
Release Number
17.20
Published
June 2022
Language
English (United States)
Last Update
2024-10-04
dita:mapPath
ncd1634149624743.ditamap
dita:ditavalPath
ruu1634160136230.ditaval
dita:id
ncd1634149624743

TD_BREUSCH_GODFREY checks for the presence of serial correlation among the residual and error terms after running a regression associated with a fitted model. With respect to regression models, it is expected that there is no serial correlation among the error terms.

The Breusch-Godfrey test determines if there is evidence of autocorrelation in the residuals of a model. Autocorrelation is correlation between the errors or residuals of a regression model meaning errors are not independent. The value of an error at a given point in time is related to the earlier error values.

TD_BREUSCH_GODFREY tests the hypothesis that there is or is not autocorrelation in the residuals. The test regresses the residuals on their lagged values for a number of lags, and then performing a chi-squared test on the residuals of the regression.

If autocorrelation is present in the residuals of a regression model, the standard errors of the estimates of the regression coefficients may be biased, and lead to incorrect inference and predictions. Using TD_BREUSCH_GODFREY in time-series analysis and econometrics, you can identify potential problems with you regression model and correct them.

The following procedure is an example of how to use TD_BREUSCH_GODFEY:
  1. Use TD_LINEAR_REGR to produce an ART containing a residual layer.
  2. Use TD_BREUSCH_GODFREY on the ART produced by the TD_LINEAR_REGR function to perform the Breusch Godfrey Statistical Test.
  3. Retrieve the TD_BREUSCH_GODFREY results from the primary layer to determine if there is serial correlation. A returned NULL_HYPOTH value of ACCEPT means no serial correlation. A returned NULL_HYPOTH value of REJECT means there is evidence of serial correlation.