Auger tries to convert Timeseries problem into classic regression problem. This is done by embedding timeseries into low-dimensional space.
This allows using Machine Learning algorithms to capture nonlinearities in timeseries as well as the power of Auger in model optimization.
In order to apply regression models to timeseries, it has to be uniform in time. Auger provides several methods for timeseries preprocessing.
This preprocessor makes observations equidistant in time.
This preprocessor fills gaps in timeseries observations using interpolation techniques.
This preprocessor embeds timeseries into lower-dimensional space which converts it to regression problem.