https://github.com/klemenjak/nilm-transferability-metrics

Metrics to assess the generalisation ability of NILM algorithms
https://github.com/klemenjak/nilm-transferability-metrics

energy-disaggregation generalisation nilmtk non-intrusive-load-monitoring overfitting transferability

Last synced: about 2 months ago
JSON representation

Metrics to assess the generalisation ability of NILM algorithms

• Host: GitHub
• URL: https://github.com/klemenjak/nilm-transferability-metrics
• Owner: klemenjak
• License: apache-2.0
• Created: 2019-12-16T13:52:31.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2019-12-16T20:41:48.000Z (about 5 years ago)
• Last Synced: 2024-10-12T15:26:38.447Z (3 months ago)
• Topics: energy-disaggregation, generalisation, nilmtk, non-intrusive-load-monitoring, overfitting, transferability
• Homepage:
• Size: 5.86 KB
• Stars: 5
• Watchers: 4
• Forks: 2
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Transferability Metrics for NILM

This repository is going to contain the implementation of our transferability metrics, which we define in our latest publication: [On Metrics to Assess the Transferability of Machine

Learning Models in Non-Intrusive Load Monitoring](https://arxiv.org/pdf/1912.06200.pdf)

PS: compatible to NILMTK!

```python

from GenLoss import *

from nilmtk.api import API

from file_handler import load_experiment

from nilmtk.disaggregate import FHMMExact, Hart85

from nilmtk_contrib import *

# load dict and execute experiment

experiment = load_experiment(experiment_ID)

api_results = API(experiment)

# assess generalisation abilities

g_loss = mean_generalization_loss(api_results)

auh = accuracy_on_unseen_houses(api_results)

exit()

```