Abstract | ||
---|---|---|
The performance of WSNs relies on the quality of the collected data. One of the main challenges is how to efficiently deliver the sensed measurements to the destination with a maximum fidelity to the probed data. Data measured by sensors require an efficient ranking measure to differentiate between normal and anomalous values. In this paper, we propose a local anomaly detector algorithm that uses Half-space trees to train and test data instances. It is based on mass ranking with difference in the way we score data and isolate anomalies. Experimental results show that our algorithm has better performance with high accuracy, low time and space complexity comparing to Local Outlier Factor and Mass estimation. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/WiMOB.2019.8923607 | 2019 INTERNATIONAL CONFERENCE ON WIRELESS AND MOBILE COMPUTING, NETWORKING AND COMMUNICATIONS (WIMOB) |
Keywords | Field | DocType |
Anomaly Detection, Binary Space Partitioning, Half-Space Tree, Mass estimation, LOF, Wireless Sensor Network | Binary space partitioning,Data mining,Local outlier factor,Anomaly detection,Fidelity,Ranking,Computer science,Test data,Detector,Wireless sensor network,Distributed computing | Conference |
ISSN | Citations | PageRank |
2160-4886 | 0 | 0.34 |
References | Authors | |
0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Alia Ghaddar | 1 | 2 | 1.71 |
Lama Darwish | 2 | 0 | 0.34 |
Fadi Yamout | 3 | 0 | 0.34 |