The Nonlinear Curve Fitter (NLFit) Tutorial Documentation

Free Demo/Evaluation Price List & Order Information System Requirements Version Comparison Feature List

The NLFit tool fits a non-linear mathematical function to user-selected data. The function can be one of many built-in non-linear fitting functions already available in Origin or can be a custom, user-defined function.

Fitting proceeds in three steps:

1. Initialization: selection of data, a function to model the data, fit-parameters, fit-parameter constraints, and number of iterations for optimization of fit parameters.
2. Calculation: Origin perfoms the math for the fitting operation.
3. Report Results: selection of which results to output and how; includes optimized parameters, and "goodness-of-fit" measures.

In the calculation step, NLFit solves a system of equations designed to minimize the aggregated point-by-point differences between the data and the selected fitting function. For non-linear fitting functions, solutions to difference equations are not typically achieved analytically, but must be obtained iteratively---this is the reason for providing a number of iterations during the initialization step. Origin uses the well-known Levenberg-Marquardt algorithm to perform the iterative computation. For more background on non-linear fitting see "Curve Fitting Made Easy", authored by a senior Origin programmer.

NLFit Features

Origin's NLFit tool is powerful, flexible, and easy to use.

• Fit all data or multiple subranges of single or multiple datasets
• Use built-in functions, or create user-defined fitting functions which can use your expressions, external DLLs or NAG Library functions
• Fit multiple datasets globally, independently, or by combining replicate measurements
• Fit with multiple independent or dependent variables
• Fit single dataset with multiple peaks using replicas of the same function
• Specify bounds and linear constraints for parameters and 12 different weighting methods
• Automatic parameter initialization for built-in functions.
• Control the Levenberg Marquardt algorithm using tolerance and number of iterations
• Customize tabular report sheets and themes for repeated use and do batch processing with custom Analysis Templates