Research and Improvement of Intrusion Detection Based on Isolated Forest and FP-Growth
Yansen Zhou, Jianquan Cui, Qi Liu · 2020
The current anomaly intrusion detection system has shortcomings such as low detection rate, high false alarm rate and poor performance in processing large amounts of data. In response to the above problems, some improvement measures are put forward for the isolated forest algorithm and the FP-Growth algorithm. The improved isolated forest algorithm considers the correlation between dimensions and makes the dimension division more reasonable for abnormal analysis. The improved FP growth algorithm reduces the time of processing a large amount of data, used for correlation analysis of abnormal data. Applying the above two improved algorithms to intrusion detection can further improve the anomaly detection performance. The results show that the false alarm rate of the joint improved algorithm is relatively reduced by 25%, and the overall detection rate is 96.24%.