Non Intrusive Load Monitoring

This project contains source code for NILM algorithms and experiments on the UK-DALE dataset.

Pre-requisites

Tensorflow
Numpy
Pandas
Matplotlib

Folder structure

Algos: Actual implementation of the models
- multi: Multi-appliance models
lib: Supporting modules
processing: Notebooks for pre-processing
experiments: different experiments run
- exp_generalization: Generalization experiments
- exp_multi_appliance: Multi-appliance experiments

Usage

Download the 2017 release of the UK-DALE dataset from https://jack-kelly.com/data/
Run the processing/processing-enhanced.ipynb file to generate data chunks for house 1
Run the processing/processing-enhanced-house-2.ipynb file to generate data chunks for house 2
Run either the experiments/sample-experiment.py file or any of the notebooks experiments in the experiments directory (ideally in a Google Colab environment)