Enhancing Non-Intrusive Load Monitoring with Features Extracted by Independent Component Analysis

2025-01-28

AI Summary

To address the challenges of scarce labeled data, low accuracy in multi-appliance concurrent disaggregation, and high model complexity in non-intrusive load monitoring (NILM), this paper proposes a lightweight end-to-end neural network architecture built upon independent component analysis (ICA). For the first time, ICA is integrated as a learnable backbone into the NILM framework, offering both physical interpretability and strong generalization capability—thereby significantly mitigating overfitting. The architecture incorporates a customized lightweight design and a novel F1-score–based multi-appliance collaborative evaluation mechanism. Experimental results on real-world data demonstrate stable, high-accuracy disaggregation for concurrent signals from ≥10 appliance types, achieving state-of-the-art F1-scores. Moreover, the proposed model reduces parameter count by 40% and decreases required training samples by approximately 60%, enabling efficient deployment under data- and resource-constrained conditions.

Technology Category

• Machine Learning: Hardware-aware ML
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

In this paper, a novel neural network architecture is proposed to address the challenges in energy disaggregation algorithms. These challenges include the limited availability of data and the complexity of disaggregating a large number of appliances operating simultaneously. The proposed model utilizes independent component analysis as the backbone of the neural network and is evaluated using the F1-score for varying numbers of appliances working concurrently. Our results demonstrate that the model is less prone to overfitting, exhibits low complexity, and effectively decomposes signals with many individual components. Furthermore, we show that the proposed model outperforms existing algorithms when applied to real-world data.

Problem

• Electricity Usage Monitoring
• Data Scarcity
• Multi-appliance Analysis

Innovation

• Neural Network Design
• Independent Component Analysis (ICA)
• Electricity Usage Data Optimization