Feature-Based Forecast Model Selection
Calculate accuracy measue based on ARIMA models
Forecast-accuracy calculation
Calculate accuracy based on MSTL
Calculate accuracy measure calculated based on neural network forecast...
Calculate accuracy measure based on random walk models
Calculate accuracy measure based on random walk with drift
Calculate accuracy measure based on snaive method
Calculate accuracy measure based on STL-AR method
Calculate accuracy measure based on TBATS
Calculate accuracy measure based on Theta method
Calculate accuracy measure based on white noise process
Autocorrelation coefficients based on seasonally differenced series
Autocorrelation-based features
build random forest classifier
Calculate features for new time series instances
Mean of MASE and sMAPE
Mean Absolute Scaled Error(MASE)
scale MASE and sMAPE by median
symmetric Mean Absolute Pecentage Error(sMAPE)
Weighted Average
Classify labels according to the FFORMS famework
identify the best forecasting method
This function is call to be inside fforms_combination
Convert multiple frequency time series into msts object
Autocorrelation coefficient at lag 1 of the residuals
calculate forecast accuracy from different forecasting methods
Combination forecast based on fforms
Function to identify models to compute combination forecast using FFOR...
Parameter estimates of Holt-Winters seasonal method
preparation of training set
function to calculate point forecast, 95% confidence intervals, foreca...
Simulate time series based on ARIMA models
Simulate time series based on ETS models
Simulate time series based on multiple seasonal decomposition
split the names of ARIMA and ETS models
STL-AR method
Unit root test statistics
A novel meta-learning framework for forecast model selection using time series features. Many applications require a large number of time series to be forecast. Providing better forecasts for these time series is important in decision and policy making. We propose a classification framework which selects forecast models based on features calculated from the time series. We call this framework FFORMS (Feature-based FORecast Model Selection). FFORMS builds a mapping that relates the features of time series to the best forecast model using a random forest. 'seer' package is the implementation of the FFORMS algorithm. For more details see our paper at <https://www.monash.edu/business/econometrics-and-business-statistics/research/publications/ebs/wp06-2018.pdf>.