Intelligent Hybrid Machine Learning Framework for Electric Line Fault Detection in Smart Homes

Abstract

Electric line fault detection in smart homes has emerged as a vital study domain within pervasive computing, wherein sensing, intelligent and communicative devices independently observe and react to environmental occurrences. Smart homes utilise embedded sensors in common objects, appliances and wearable gadgets to gather data regarding human activities and environmental conditions. To safeguard data privacy and enhance model efficacy, preprocessing methods including data anonymisation, feature vector extraction and label analysis are employed. This paper proposes a novel CEBAS-ANFIS model for intelligent fault identification and prediction. The methodology integrates an advanced BAS swarm intelligence optimisation technique with machine learning and ANFIS to refine model parameters and enhance prediction efficacy. The improved BAS algorithm rectifies the shortcomings of the traditional BAS approach and augments the system's overall learning capacity. Performance evaluation was conducted using a wide range of benchmark functions and comparative analyses against several state-of-the-art methods on power quality fault detection datasets. The experimental results indicate that the proposed CEBAS-ANFIS approach much surpasses current methodologies, with a precision of 94.36%. The findings validate that the CEBAS-ANFIS framework provides a dependable, precise and effective solution for Electric Line Fault Detection in Smart Homes, facilitating the advancement of safer and more intelligent living environments.