seer1.1.8 package

Feature-Based Forecast Model Selection

accuracy_arima

Calculate accuracy measue based on ARIMA models

accuracy_ets

Forecast-accuracy calculation

accuracy_mstl

Calculate accuracy based on MSTL

accuracy_nn

Calculate accuracy measure calculated based on neural network forecast...

accuracy_rw

Calculate accuracy measure based on random walk models

accuracy_rwd

Calculate accuracy measure based on random walk with drift

accuracy_snaive

Calculate accuracy measure based on snaive method

accuracy_stlar

Calculate accuracy measure based on STL-AR method

accuracy_tbats

Calculate accuracy measure based on TBATS

accuracy_theta

Calculate accuracy measure based on Theta method

accuracy_wn

Calculate accuracy measure based on white noise process

acf_seasonalDiff

Autocorrelation coefficients based on seasonally differenced series

acf5

Autocorrelation-based features

build_rf

build random forest classifier

cal_features

Calculate features for new time series instances

cal_m4measures

Mean of MASE and sMAPE

cal_MASE

Mean Absolute Scaled Error(MASE)

cal_medianscaled

scale MASE and sMAPE by median

cal_sMAPE

symmetric Mean Absolute Pecentage Error(sMAPE)

cal_WA

Weighted Average

classify_labels

Classify labels according to the FFORMS famework

classlabel

identify the best forecasting method

combination_forecast_inside

This function is call to be inside fforms_combination

convert_msts

Convert multiple frequency time series into msts object

e_acf1

Autocorrelation coefficient at lag 1 of the residuals

fcast_accuracy

calculate forecast accuracy from different forecasting methods

fforms_combinationforecast

Combination forecast based on fforms

fforms_ensemble

Function to identify models to compute combination forecast using FFOR...

holtWinter_parameters

Parameter estimates of Holt-Winters seasonal method

prepare_trainingset

preparation of training set

rf_forecast

function to calculate point forecast, 95% confidence intervals, foreca...

sim_arimabased

Simulate time series based on ARIMA models

sim_etsbased

Simulate time series based on ETS models

sim_mstlbased

Simulate time series based on multiple seasonal decomposition

split_names

split the names of ARIMA and ETS models

stlar

STL-AR method

unitroot

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>.

  • Maintainer: Thiyanga Talagala
  • License: GPL-3
  • Last published: 2022-10-01