nilmtk/nilmtk

What it solves

It provides a standardized data and evaluation layer for Non-Intrusive Load Monitoring (NILM), simplifying how researchers handle energy datasets, preprocess data, and measure the accuracy of energy disaggregation models.

How it works

NILMTK acts as a core toolkit that converts various public energy datasets into a common HDF5 format. It provides a Python API for lazy access to buildings, meters, and appliances, and handles the heavy lifting of resampling, alignment, and calculating standard NILM accuracy and energy metrics.

Who it’s for

Researchers and developers working on energy disaggregation, specifically those focused on data preparation, meter abstractions, and performance evaluation.

Highlights

• Dataset Converters: Built-in tools to convert public energy datasets into a unified format.
• Standardized Metrics: Implements standard NILM accuracy and energy metrics for consistent evaluation.
• Data Management: Provides lazy access to buildings, meters, and appliances to handle large datasets efficiently.
• Reference Algorithms: Includes classical reference disaggregators and baseline utilities for initial experimentation.