One Class Support Vector Machine for Anomaly Detection in the Communication Network Performance Data

摘要

Anomaly detection refers to the automatic identification of the abnormal behaviors from a large amount of normal data. Anomaly detection is more and more required in the communication network due to the increasing number of the unauthorized activities occurring in the network. This paper presents a method based on one class support vector machine (OCSVM) to detect the network anomalies. The communication network performance data are used for the investigation and the raw data are firstly preprocessed in order to produce the vector sets required by the OCSVM algorithm. The training data are used to train the OCSVM anomaly detector, and the trained detector is applied on the test data to detect the anomalies. In addition, the results are compared with the results obtained from the rule-based system which is currently used in the communication network. The algorithm shows the promising performance on the network anomaly detection and provides a great reduction on the volume of the alarms than the rule-based system.