https://github.com/ihomelab/dnn4nilm_overview
Overview of NILM works employing Deep Neural Networks on low frequency data
https://github.com/ihomelab/dnn4nilm_overview
energy-disaggregation nilm
Last synced: 6 months ago
JSON representation
Overview of NILM works employing Deep Neural Networks on low frequency data
- Host: GitHub
- URL: https://github.com/ihomelab/dnn4nilm_overview
- Owner: ihomelab
- License: mit
- Created: 2019-10-10T13:50:38.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2021-04-23T11:28:51.000Z (over 4 years ago)
- Last Synced: 2024-08-02T15:54:10.206Z (about 1 year ago)
- Topics: energy-disaggregation, nilm
- Language: Jupyter Notebook
- Size: 771 KB
- Stars: 12
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Review on Deep Neural Networks applied to Low-Frequency NILM
This repo contains data and code that has been used for the publication
"Review on Deep Neural Networks applied to Low-Frequency NILM" submitted @ MDPI
Energies [doi.org/10.3390/en14092390](https://doi.org/10.3390/en14092390).
This work is a considerable extension of the presentation "DNN for NILM on low
frequency Data" that has been done at the NILM workshop 2019. You can find the
corresponding presentation
[here](https://www.youtube.com/watch?v=010fawyCOCs&list=PLJrF-gxa0ImryGeNtil-s9zPJOaV4w-Vy&index=11)
Content:
* `DNN-NILM_Publication-List.xlsx` contains the list of the DNN-NILM
publications that have been reviewed in the mentioned publication. It
corresponds with minor differences in columns and nomenclature to table 2 in
the publication and is provided to allow for easy searching and filtering.
Abbreviations are explained in the publication.
* `Visualize_MAE.ipynb` and `Visualize_F1.ipynb` are the jupyter notebooks that
have been used to generate the visualizations in the paper, i.e. figures 3
and 4. Please be aware that citation numbering might have changed in the
final publication.
* `DNN-NILM_low-freq_Performance.xlsx` contains the list of metrics extracted
from the reviewed publications. Publications that did
* not report metrics,
* report metrics other than F_1-score or MAE or
* not report metrics according to the relevant evaluation scenario
might not appear in the list. The file is the basis for the figures generated
with the jupyter notebooks. Some explanations on the columns can be found in
the tab `Explanations`. Please do not expect that *all* columns are filled up
consistently.
In case you are an author of one of the publications and feel that erroneous
information has been compiled in our list, do either contact
patrick.huber@hslu.ch or open a pull request with your suggested changes. We
will appreciate your feedback!