Improving intrusion detection using a novel normalization method along with the use of harmony search algorithm for feature selection

In recent years, due to increasing rate of traffic in computer networks as an issue of concern for the community of security researchers, more accurate and faster intrusion detection algorithms are needed to be developed. Thereafter, the advances in terms of feature selection using genetic algorithm and preprocessing methods have paved the way to detect intrusive activities. In this regard, the attempts were made in this study to present a novel method in transferring character data into numerical data to make them suitable to be used in harmony search-support vector machine (HS-SVM). To this end, the NSL-KDD dataset is utilized to present the effectiveness and accuracy of HS-SVM classification. The findings of the present study suggest that the proposed model yielded better performance in terms of speed and accuracy of detecting intrusion compared to other studied preprocessing methods.