Probabilistic Load Forecasting based on Adaptive Online Learning (APLF)

This repository contains code for the paper Probabilistic Load Forecasting based on Adaptive Online Learning [1]. We use the implementation details described in the paper.

Implementation of the method

APLF folder contains the Python file, a Jupyter notebook and a Matlab folder that contains all the Matlab scripts required to execute the method:

APLF.m is the main file.
initialize.m function inizializes model parameters.
prediction.m function obtain load forecasts and probabilistic load forecasts in form of mean and standard deviation of a Gaussian density function.
test.m function quantifies the prediction errors RMSE and MAPE.
update_model.m function updates the model for each new training sample.
update_parameters.m function updates model parameters.

Data

We use 7 publicly available datasets corresponding with regions that have different sizes. The datasets are:

Load demand in Belgium from 2017-2019 made available by Elia group.
Load demand in New England from 2003-2014 made available by ISO-NE organization.
Global Energy forecasting Competition 2012 dataset from 2004-2007.
Global Energy Forecasting Competition 2014 dataset from 2005-2011.
Load demand in Dayton from 2004-2016 made available by PJM interconnection.
Load demand for 400 buildings in New South Wales from 2013 made available by the Australian Government.
Load demand for 100 buildings in New South Wales from 2013 made available by the Australian Government.

We save the data in .mat files that contain a struct with following fields:

Hourly load time series
Temperature time series
Date and hour or timestamp when the load is measure

Installation

git clone https://github.com/VeronicaAlvarez/online-probabilistic-load-forecasting

Running python code:

cd online-probabilistic-load-forecasting\APLF
python APLF.py

Test case

We display in this reposity an example for a dataset that contains load data of 400 buildings. Example folder includes more details of the dataset, commands to execute the code, and results.

Support and Author

Verónica Álvarez Castro

valvarez@bcamath.org

License

Load-forecasting-IEEE-TPWRS-2020 carries a MIT license.

Citation

If you find useful the code in your research, please include explicit mention of our work in your publication with the following corresponding entry in your bibliography:

[1] V. Alvarez, S. Mazuelas, J.A. Lozano. "Probabilistic Load Forecasting based on Adaptive Online Learning," IEEE-Transactions on Power Systems. 2021.

The corresponding BiBTeX citation is given below:

@article{AlvMazLoz:21,
 title={Probabilistic Load Forecasting based on Adaptive Online Learning},
 author={Ver\'{o}nica Alvarez and Santiago Mazuelas and Jos\'{e} A. Lozano},
 journal={IEEE Transactions on Power Systems},
 year={2021},
 volume={36},
 number={4},
 month= {Jul.},
 pages={3668 -- 3680}
}

MachineLearningBCAM/Load-forecasting-IEEE-TPWRS-2020