Interpolation results and parameter ordering
The order of parameters associated with grid Functional data can have an affect on interpolation results.
If you are using grid Functional data where one or more of the parameters use cubic spline interpolation, then parameter order may have an effect on the interpolation results. Note that parameter order has no effect if only linear interpolation is being used, or in series functional data.
Parameter order is important because gridded graphs are interpolated one parameter at a time. So, for a graph with 3 parameters (strain, cycles and time), to interpolate to a single point:
- First constrain a parameter (strain = 0.1%, say) and create a new graph that only has the parameters cycles and time.
- Then constrain another parameter (cycles = 1,000,000, say) to get a graph that only has one parameter.
- Finally, constrain the last parameter (time) to get a point value.
If a non-linear interpolation strategy is used for one of those steps, the value of the interpolated point can vary depending upon the order in which the parameters are interpolated.
Note that each piece of functional data can have the parameters stored in a different order. If version control is being used, then each version of the data can have a different order.
The order of parameters on functional data is set when you view the data in MI Viewer, or when data is imported by the Excel or Text importer. In MI Viewer, the order of the parameters is controlled by the order of the parameter columns. Grid functional data has the following format. In this example, Parameter 1 will be first, followed by Parameter 2, etc.
| Column | Contains |
|---|---|
| 1 | Value (low) |
| 2 | Value (high) |
| 3 | Quality rating |
| 4 | Parameter 1 |
| 5 | Parameter 2 |
| 6 | Parameter 3 |
| ... | Parameter n |
To check the order of parameters, open Graph Tools and click View All Data. The order that the parameters appear in the columns is the order they are stored in the functional data.