Automatic Model Selection and Prediction for Univariate Time Series
Creates additional dates and values when NA where removed and the TS i...
Determine best algorithm
Determines the decimal frequency of a time series from a character str...
Fit BATS algorithm and make the prediction
Fit ETS algorithm and make the prediction
Custom (internal) function for MAE
Make predictions with selected algorithms
Fit prophet algorithm and make the prediction
Custom (internal) function for RMSE
Fit SARIMA algorithm and make the prediction
Fit short term algorithm and make the prediction
Fit STLM algorithm and make the prediction
Fit TBATS algorithm and make the prediction
Format 2 vectors in a proper object usable by all algorithms
Demo graphical user interface
Offers a set of functions to easily make predictions for univariate time series. 'autoTS' is a wrapper of existing functions of the 'forecast' and 'prophet' packages, harmonising their outputs in tidy dataframes and using default values for each. The core function getBestModel() allows the user to effortlessly benchmark seven algorithms along with a bagged estimator to identify which one performs the best for a given time series.