Applying Semi-supervised Cluster Algorithm for Anomaly Detection

Most current anomaly detection systems employ supervised methods or unsupervised methods. Supervised methods rely on labeled training data, however, in practice, this training data is typically expensive to produce. In contrast, unsupervised method can work without the need for massive sets of pre-labeled training data, but its accuracy is quite low. This paper put forward a semi-supervised cluster algorithm to detect anomalies in network connections, we evaluate our method by performing experiments over network records from the KDD CUP99 data set. The result shows that the feasibility of this method is much better than the others.