This folder contains the original data from the EEM 2017 Wind Power Forecasting Competition and the code use by the wining team, “p9”, Jethro Browell and Ciaran Gilbert. It's intended purpose is to enable scrutiny of our methodology and to serve as a benchmark for future improvement. This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ or send a letter to Creative Commons, PO Box 1866, Mountain View, CA 94042, USA. Attibution: Jethro Browell, Ciaran Gilbert If you use this data and code, please cite the ditital object identifier: DOI: 10.15129/01f10e89-3a33-4437-8b2e-9b53c70dad4f This file contains a description of the R scripts used for model selection and estimation. For more detail, please see the accompanying paper, and cite if used: J. Browell, C. Gilbert, "Cluster-based regime-switching AR for the EEM 2017 Wind Power Forecasting Competition," 14th International Conference on the European Energy Market (EEM), Dresden, Germany, 2017. For further details of the competitoin see: http://eem2017.com/program/forecast-competition FILES meteorological_data.csv and wind_power_generation_data.csv: Training data provided by the competition organisers. NB: Access to power data has been restricted - we are working with the data owners to try and have this restriction lifted. CompData.zip: Competition period in-put data, results published by the competition organisers. NB: Restricted as above. CompResults.zip: Results and ranking as published on a roling basis during the competiton period. p9_Entries.zip: The entries submitted by “p9” during the competition period. CV_TimeSeriesModels.R: Implements and evaluates a range of time series models using hold-out cross-validation. Each month from Feb-Dec is “held-out” in turn and used for out-of-sample forecast evaluation. CV_NumberOfClusters.R: Cross-validation as above to determine the optimal number of weather regimes/clusters. TrainingScript_CompModels_v2.R: Fits best performing models to entire training dataset and saves them for use during competition period. MakeEntries.R: Loads competition-period input data and outputs a .csv containing a forecast for submission to the competiton. CONTACT INFORMATION Jethro Browell Ciaran Gilbert