ملخص البحث :
Even security specialists find it challenging to monitor the complex interconnections of
computers and network devices brought about by the expansion of the internet over the past ten years.
Network security has grown to be a major problem as personal computers have become faster and highspeed
internet has become more widely accessible. It is extremely difficult to create intrusion detection
systems that can manage massive amounts of data, especially in terms of system construction time. This
work suggests a preprocessing feature selection strategy that creates subsets of pertinent characteristics to
ease model construction in order to overcome this difficulty. The suggested model uses the information
gain method to improve accuracy while classifying network data using the Random Forest algorithm. Using
the NSL-KDD reference dataset, the suggested model's efficacy is assessed. Several measures are used to
determine how well it performs. According on empirical findings, the recommended model outperforms
existing algorithms in terms of performance measures. It offers a contrast. Overall, the proposed
methodology has a great deal of promise to enhance large data intrusion detection systems' functionality.
Keywords: Cybersecurity, Random Forest Algorithm, Big Data, Intrusion Detection System, Performance Analysis,
Threat Detection, Anomaly Detection
-
سنة النشر : 2024
-
تصنيف البحث : scopus
- تحميل