Visualization in big data analytics: Applying hidden Markov models to big data: an appliance modeling case study
Type
Supervisor
Subject
Date
2026-01-17
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
Accurate energy consumption data and forecasts at the individual appliance level empower users to make informed decisions, reduce electricity costs, engage in targeted energy conservation, and promote sustainability. This research focuses on the intersection of the Reference Energy Disaggregation Dataset (REDD) and Hidden Markov Models (HMMs) for electric household appliances. The study specifically models the behavior of a two-state electric lamp appliance using a state-based HMM. Furthermore, an adaptive model is developed, incorporating three lamps with the same Finite State Machine (FSM), to predict the behavior of an unknown fourth lamp. The accuracy of these models is evaluated over different time spans (9, 4.5, and 2.25 hours) by generating a Viterbi sequence state. Accuracy measures validate the results, indicating that the best-performing single-variant model achieves 100% accuracy, while the best-performing three-variant model achieves 97.08%. The proposed system demonstrates its functionality and p
Department
Publisher
Sponsor
Effat University
Copyright
Book title
Mathematical Modeling for Big Data Analytics