This tool provides a convenient way to perform curve fitting. With the data plotted, simply open the tool, type the desired equation, specify initial parameters and generate your report. You do not need to even save or create a fitting function.
The tool only supports one independent variable (x) and one dependent variable (y).
Download the file Simple Fit.opx, and then drag-and-drop onto the Origin workspace. An icon will appear in the Apps Gallery window.
- Create a graph of the XY data with optional error bars, and click the Simple Fit app icon.
- Choose one of the fitting types: linear, polynomial, or nonlinear.
- If error bar is present in the graph, you can optionally perform a weighted fitting by checking the Weighted Fit check box.
- If there are multiple plots in the graph, you can optionally fit all plots (individually) by checking the Fit All Curves in the Graph check box.
- To display the 95% confidence/prediction band, check the corresponding check box.
- When performing nonlinear fitting:
- Define the fitting function:
- To use an existing function, check the Use Existing Function check box and select desired function.
- To use a new function, uncheck the Use Existing Function, enter the function definition into the "y(x)" edit box, using "x" as independent variable.
- Specify or change desired initial parameter values.
- Click the 1 Iter. button to perform one iteration at a time, or the Fit button to perform iterations until the tolerance is reached or the maximum number of iterations (400) has reached. Before each operation, you can change or fix parameter values.
The last column about statistics can be changed to show one of Residual Sum of Squares, Reduced Chi-Sqr, R-Square and Adj. R-Square by context menu.
- To save the fitting function, click the Save as... button. The fitting function will be saved in User Defined category.
If you do not save the function, it will not be available for selection next time the tool used.
NOTE: When performing weighted fitting, the error values reported on the fit parameters are NOT scaled by the sqrt(reduced chi-square) value.