A Novel Probabilistic Hybrid Model to Detect Anomaly in Smart Homes
Anomaly detection in smart homes provides support to enhance the health and safety of people who live alone. Compared to the previous studies done on this topic, less attention has been given to hybrid methods. This paper presents a novel probabilistic hybrid model to detect anomaly in the smart home. It detects anomaly in two steps. First, it employs various algorithms belonging to different algorithm type categories for anomaly detection from sensory data. Then, a Bayesian network is trained on a dataset consist of test results of employed algorithms and the actual results. It hybridizes results of used methods and adjusts their impact on the final anomaly decision. Finally, abnormal events were detected through calculating the probability of actual result’s node given test results of other network’s nodes. Furthermore, we use the structure of the trained network to indicate conditional independence relationships between different algorithms. Experimental evaluations of a real dataset indicate the effectiveness of the proposed method. It reduced false positives and increased true positives.
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles published by TSP are under an Open Access license, which means all articles published by TSP are accessible online free of charge and as free of technical and legal barriers to everyone. Published materials can be re-used if properly acknowledged and cited Open Access publication is supported by the authors' institutes or research funding agencies by payment of a comparatively low Article Processing Charge (APC) for accepted articles.