Intrusion Detection Systems (IDS) are essential tools for securing today's network infrastructures against malicious activities and unauthorized access. Compared with other attack categories, Remote-to-Local (R2L) and User-to-Root (U2R) are still very difficult to attack because they are infrequent and similar to regular network operations. This work proposes an optimized intrusion detection system framework based on DNA for better detection capability of R2L and U2R attacks in the NSL-KDD dataset. The proposed framework includes the preprocessing, transforming features, improved DNA encoding, feature fusion, and attack classification in the IDS pipeline. The DNA encoding mechanism converts network traffic features into DNA-inspired features to enhance feature expressiveness and attack discrimination. Experimental evaluation proves the proposed framework to be effective in the separation of malicious traffic from normal traffic in the network. The results of the confusion matrix analysis and visualization also support the effectiveness of the proposed framework in dealing with the difficult intrusion category. The results suggest that feature transformation is an effective way to enhance the capabilities of intrusion detection systems, especially against behaviorally hidden attacks such as R2L and U2R.

https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/https://fyp.coie-nahrain.edu.iq/index.php/supervisor-submission/

2. The research survey is not enough. The articles listed in References are not state-of-the-art. (Most of them are out of date. For example, the papers published within 2 years are used to calculate CiteScore in SCOPUS.) The authors should survey past studies published after 2025. Furthermore, the authors should justify the effectiveness of the proposed method by comparing with state-of-the-art methods.
3. Overall, the authors have made a good attempt. However, the authors' proposed method does not adequately describe their data. The results are not supported by any theoretical/mathematical reasons. Readers will fail to understand the scientific contribution of this research. The authors should justify the effectiveness of the proposed technique theoretically.