You can use this analysis object to calculate various regressions for a data set.
If you know the functional relationship between the X and the Y components of a signal, you can replace the data with a formula that approximates the data as precisely as possible. To do this, specify a model function where still unknown coefficients are used, which are to be determined such that the square error between the raw data and its mathematical representation is as low as possible. This goodness-of-fit measure, also known as χ² (Chi-squared), is equal to the sum of the squares of all deviations of the approximated model function of the data.
FlexPro allows you to calculate the following regressions:
Description |
Model Function |
|---|---|
Linear Regression |
f(x) = a + b * x |
Logarithmic Regression |
f(x) = a * Log10(b * x) |
Exponential Regression |
f(x) = a * exp(b * x) |
When calculating the regression, both coefficients a and b are determined.
Dynamic and Static Regression
The regression offers you two alternatives for calculating the coefficients: You can calculate the factors once and enter them as static into the regression function, or use a dynamic calculation for the factors. In this case, the factors are always re-calculated when the source data is changed.
Regression Function and Coefficients
You can choose whether would like the result to be the regression function, i.e., a formula in which the coefficients were substituted by the calculated values, or the regression coefficients as data series with three values. The third value in the data series is the goodness-of-fit measure χ².
Evaluating the Regression Function
While the data set only contains discrete points on the characteristic curve, you can freely determine the X values for which the function is to be evaluated for the regression calculated. For this, on the Data tab, select the option Use a different data set as independent variable X and specify a data set for the X values. The regression is also suited for "re-sampling" signals.
FPScript Functions Used
See Also
* This analysis object is not available in FlexPro View.