This repository provides an easy to use working implementation of Latent Bayesian Melding for NILM (energy disaggregation). For more details refer to our NIPS 2015 paper: Latent Bayesian melding for integrating individual and population models (http://papers.nips.cc/paper/5756-latent-bayesian-melding-for-integrating-individual-and-population-models; or http://arxiv.org/abs/1510.09130).

Mosek academic license (free)

I found that the following Mosek version works: mosek-7.1.15-py2.7. I have tried higher versions, but it did not work.

To install Mosek, fully follow the instruction from here: http://docs.mosek.com/7.1/toolsinstall/Linux_UNIX_installation_instructions.html.

You can find the Mosek package in this repository, but you need a license to use it.

Just run the script: demo_latentBayesianMelding.py

Contact: You may contact the first author at: mingjun.zhong@gmail.com or mzhong@lincoln.ac.uk