Parameters | Predictive Modeling | Root Mean Square Error Convergence Tolerance

Root Mean Square Error Convergence Tolerance
Use this option to specify a convergence tolerance (between 0 and 100) for the Buckley-James iterations.
The tolerance is for the coefficient of variation of the root mean square error (CV(RMSE)) between predicted and observed values. This value is computed by dividing the usual root mean square error by the mean of the dependent variable.
Buckley-James iterations stop when CV(RMSE) drops below the specified tolerance.
To Specify a RMSE Convergence Tolerance:
*