Voluminous data with high velocity and variety have resulted in deceiving the security of internet and intranet facilities. The threats are either having some patterns or lack any definite patterns. Therefore, the data arriving at the network have number of features and wide variety of patterns. Firstly, the number of patterns needs to be reduced and then the filtered set of patterns could be used for detecting unknown threats. This paper presents an approach for developing an Intrusion Detection System (IDS) with the help of Principal Component Analysis (PCA) and machine learning algorithms in WEKA environment. The approach yields better performance by making the detection more effective. The results show highertrue positive and lower false positive ratesin comparison to the existing methods.
A Proposed Machine Learning based Scheme for Intrusion Detection
2018-03-01
2414312 byte
Conference paper
Electronic Resource
English
A Hybrid Machine Learning Based Intrusion Detection System for MIL-STD-1553
DOAJ | 2024
|