What is 'Durbin-h' in SAP SCM-APO-FCS - Demand Planning?

Durbin-h - Overview

• Component: SCM-APO-FCS

• Component Name: Demand Planning

• Description: The Durbin-h statistic tests for autocorrelation in time series where independent variables are lagged by one or more periods. If Durbin-h is equal to or greater than 1.96, it is likely that autocorrelation exists. The Durbin-h test is suitable for large samples; that is, for samples of 100 or more.

Durbin-h - Details

• 
  Key Concepts: 
  Durbin-h is a statistical forecasting method used in SAP's Supply Chain Management - Advanced Planning and Optimization - Forecasting and Simulation (SCM-APO-FCS) Demand Planning component. It is a variant of the Durbin-Watson test, which is used to detect autocorrelation in time series data. The Durbin-h test is used to detect autocorrelation in the errors of a regression model. 
  
  How to use it: 
  The Durbin-h test can be used to determine whether a regression model is appropriate for forecasting demand. If the Durbin-h statistic is greater than 1, then the errors of the regression model are autocorrelated, which means that the model is not suitable for forecasting. If the Durbin-h statistic is less than 1, then the errors of the regression model are not autocorrelated, which means that the model can be used for forecasting. 
  
  Tips & Tricks: 
  When using the Durbin-h test, it is important to remember that it only detects autocorrelation in the errors of a regression model. It does not detect autocorrelation in the data itself. Therefore, it is important to use other methods such as the Durbin-Watson test to detect autocorrelation in the data before using the Durbin-h test. 
  
  Related Information: 
  The Durbin-h test is related to other statistical tests such as the Durbin-Watson test and the Breusch-Godfrey test. These tests can be used to detect autocorrelation in time series