ifeel 1.1.0

A python package for Interpretable Feature Extraction of Electricity Loads (IFEEL)

Interpretable Feature Extraction of Electricity Loads (IFEEL)

A python package for Interpretable Feature Extraction of Electricity Loads (IFEEL)

• IFEEL on GitHub 🔗

• IFEEL on PyPI 🔗

📌 Illustration:

Note: If the picture fails to load, please click here.

⚙ Installation:

You can use pip to easily install IFEEL with:

pip install ifeel

More info about pip can be found here .

🤖 Developer info:

• Package title: Interpretable Feature Extraction of Electricity Load (IFEEL)

• Authors: Maomao Hu, Dongjiao Ge, David Wallom

• Organization: Oxford e-Research Center, Department of Engineering Science, University of Oxford

• Contact info: [e-mail]

• Development time: Oct 2020

• Acknowledgement: This work was financially supported by the UK Engineering and Physical Sciences Research Council (EPSRC) under grant (EP/S030131/1) of AMIDINE.

💬 About IFEEL:

(1) This Python package (i.e., IFEEL) aims to help energy data analysts to readily extract interpretable features of daily electricity profiles from a physical perspective. The extracted features can be applied for further feature-based machine learning purposes, including feature-based PCA, clustering, classification, and regression.

(2) Two PY files (.py) are included in the IFEEL package, including ifeel_transformation.py and ifeel_extraction.py.

(3) Two types of features can be extracted by using this package: 13 global features (GFs) and 8 peak-period features (PFs).

(4) The global features are extracted based on raw time-series data, while the peak-period features are extracted based on symbolic representation of time series. The feature extraction process is performed by calling the functions in ifeel_extraction.py.

(5) For fast peak-period feature extraction, Symbolic Aggregate approXimation (SAX) representation is first used to transform the time-series numerical patterns into alphabetical words. The feature transformation process is performed by calling the functions in ifeel_transformation.py. More details about SAX approach can be found in Ref [2] and Ref [3].

🔈 Notes:

(1) To successfully run the IFEEL, the following Python data analysis libraries need to be installed in advance: Numpy, Scipy, and Pandas.

(2) A Demo can be found in the installed IFEEL package or here. The dataset used in the Demo can be downloaded here.

(3) The Demo has been tested on Python 3.7.7.

📚 References

[1] Hu M, Ge D, Telford R, Stephen B, Wallom, B. Classification and characterization of intra-day load curves of PV and Non-PV households using interpretable feature extraction and feature-based clustering. Energy.(Under review)

[2] Lin J, Keogh E, Wei L, Lonardi S. Experiencing SAX: a novel symbolic representation of time series. Data Mining and Knowledge Discovery. 2007;15:107-44.

[3] Keogh E, Lin J, Fu A. HOT SAX: efficiently finding the most unusual time series subsequence. 5TH IEEE International Conference on Data Mining (ICDM'05). 2005. p8.