Common Problems In Curve FittingΒΆ
Single Outlier At End Point |
Outliers are often caused by manual errors in recording experimental data. |
Single Outlier At Mid Point |
Outliers are often caused by manual errors in recording experimental data. |
Fitting Parallel Data |
This problem is often caused by environmental changes during data collection, such as temperature changes on different days when making multiple data collection runs. |
Data With A Large Step |
This problem is often caused by environmental changes during data collection, such as a temperature change during a lunch break. |
Data With A Poorly Defined Region |
This problem can be mitigated by taking additional data in the region that is poorly defined. |
Equation Missing An Offset |
This illustrates the effect of fitting data with an offset to an equation that does not have one. This problem can be caused by experimental equipment introducing bias (such as a DC offset) during data acquisition. Fitting the data to an equation with an offset will reveal the bias. |
Data Scatter Over Entire Range |
The effect of data scatter (noise) can be reduced by increasing the total number of data points. |
Fitting Random Data |
This illustrates the effect of fitting completely random data that has no relationship of any kind. |
Based on the pyeq3 CommonProblems example at BitBucket written in Python 3.
E-mail zunzun@zunzun.com for questions or suggestions.