This model uses cubic spline interpolation to fill in gaps in a time series. The time series in this model contains 2 kinds of missing data: 1. missing values along with their dates are omitted and 2. missing values are replaced with a value of -99. In this example, an external DLL is used to perform cublic spline interpolation, which relies on the The GNU Scientific Library (GSL).
This can be an effective way to handle missing data in a time series when the missing values do not span over more than a few time points when the values are highly variable. In the example shown below, nearly the entire month of August is missing but the interpolation is still reasonable because it is during a period of smaller changes. You should check the interpolation results before using them in your model.
This model requires that you also save the DLL in the same directory as the model file before running it. Please make sure you download both files from the links below.
Please sign in to leave a comment.