How To Fix RSAN_RT322 - Tolerance < 0.02: Correlation coefficients saved, not regression coeffs

RSAN_RT322 - Overview

• Message type: E = Error

• Message class: RSAN_RT - Messages for Runtime Analysis

• Message number: 322

• Message text: Tolerance < 0.02: Correlation coefficients saved, not regression coeffs

MESS SHORT TEXT:
Tolerance < 0.02: Correlation coefficients saved, not regression coeffs

What causes this issue?

The tolerance for one of the regression coefficients is smaller than
0.02.

System Response

A tolerance of below 0.02 means strong collinearity and thereby high
standard errors for the regression coefficients.
For this reason, bivariate correlation coefficients (which are not
influenced by collinearity) are saved instead of the regression
coefficients.

How to fix this error?

The system issues an error message and will not allow you to continue with this transaction until the error is resolved.

Procedure for System Administrators

Error message extract from SAP system. Copyright SAP SE.

Smart SAP Assistant

• What is the cause and solution for SAP error message RSAN_RT322 - Tolerance < 0.02: Correlation coefficients saved, not regression coeffs ?

  The SAP error message RSAN_RT322 indicates that during the execution of a regression analysis in SAP BW (Business Warehouse), the tolerance for the correlation coefficients was set below the threshold of 0.02. As a result, the regression coefficients could not be saved, and only the correlation coefficients were retained.
  
  Cause:
  
  Low Tolerance Level: The error is triggered when the correlation coefficients of the variables in the regression analysis are too low, indicating that the variables may not have a strong linear relationship. The default tolerance level is set to 0.02, meaning that if the correlation coefficient is below this value, the regression coefficients will not be saved.
  Data Quality Issues: The underlying data may have issues such as insufficient variability, outliers, or noise, which can lead to low correlation between the independent and dependent variables.
  Model Specification: The model may not be correctly specified, meaning that the chosen independent variables may not be appropriate