Scholarlink

<p>Non-Intrusive Load Monitoring (NILM), also referred to as energy disaggregation, is the task of estimating a device level energy consumption using the aggregated energy consumption of different devices at one single measurement point without installing meters on each individual device. NILM can be formulated as a source separation problem where the aggregated signal is expressed as linear combination of basis vectors in a matrix factorization framework. In this paper, we propose a novel Bayesian Non Negative Matrix Factorization for energy disaggregation. The model achieves superior performance by imposing sparcity on the activation matrix using Dirichlet priors. To estimate the parameters of the model, variational Bayesian inference is used. A lower bound approximation for the objective function is used to find an analytically tractable solution for the model. We evaluate the model and show its merits with three data sets: REDD, AMPds and IRISE, and with multiple experimental setups.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Energy disaggregation via Bayesian non-negative matrix factorization with sum-to-k constraint

  • 期刊论文

  • 发表日期：

    2025年1月16日

查看全文
  • Oumayma Dalhoumi,
  • Manar Amayri,
  • Nizar Bouguila

摘要

Non-Intrusive Load Monitoring (NILM), also referred to as energy disaggregation, is the task of estimating a device level energy consumption using the aggregated energy consumption of different devices at one single measurement point without installing meters on each individual device. NILM can be formulated as a source separation problem where the aggregated signal is expressed as linear combination of basis vectors in a matrix factorization framework. In this paper, we propose a novel Bayesian Non Negative Matrix Factorization for energy disaggregation. The model achieves superior performance by imposing sparcity on the activation matrix using Dirichlet priors. To estimate the parameters of the model, variational Bayesian inference is used. A lower bound approximation for the objective function is used to find an analytically tractable solution for the model. We evaluate the model and show its merits with three data sets: REDD, AMPds and IRISE, and with multiple experimental setups.