Intrusion detection system model using Fuzzy Kernel C-Means and Laplacian Score feature selection

Abstract

Nowadays, where the technology dominates every field and activity like transaction, learning activities, private or corporate data storage, and so on are things that we have to look in detail especially in terms of security both in data storage and technology utilization. The trust of the security itself can be represented by the accuracy on model we used but the question is “Which model is good for the security?”, that question can be answered by the research or trial of models either by combining the models or construct the new one. Therefore, in this paper we tried an intrusion detection system model by combining Fuzzy Kernel C-Means method as a classifier and Laplacian Score method as a feature selection applied on KDD Cup 99 Data Set. As a result, we will see the comparison of the accuracy of intrusion detection system (IDS) model using Fuzzy Kernel C-means with and without Laplacian Score Feature Selection for the same data and features that resulted 60% and 67.06%, respectively.