Now 5234 visitors
Today:0 Yesterday:0
Total: 147759 212S 50P 4R
2026-10-20, Week 43
Member Login
TACT Journal Homepage
Call for Paper
Committee
Find My Paper
Author Homepage
Statistics & History
Welcome Message
Paper Procedure
FAQ
Paper Archives
Outstanding Papers
Program & Proceedings
Presentation Platform
Hotel & Travel Info
Photo Gallery
Scheduler Login
Seminar
Archives Login
Sponsors


















Online Q&A Session
Paper Number Author:
Paper Title
Keyword
Q&A Number **Select your question number !!
Questioner eMail:
Question
Answer
by
Author
Save Q&A

* Edit or answer any Q&A by selecting Q&A number Hyper Link below + Write button (Save)
ICACT20260118 Question.1
Questioner: hripsime.hovhannisyan@rau.am    2026-02-09 오전 11:28:46 		ICACT20260118 Answer.1
Answer by Auhor sudhakar@ccet.ac.in   2026-02-09 오전 11:28:46
Chrome Click!!
When you added the Anomaly Override layer, how did it affect the false positive rate? Specifically, did it increase or decrease the number of false alarms, and by what percentage compared to the standard ensemble model? That is a very insightful question. In our framework, the addition of the Anomaly Override layer was primarily designed to minimize critical false negatives, which we consider the highest risk in malware detection. Regarding the False Positive Rate, we observed that it remained stable and exceptionally low. As you can see from our confusion matrix on Slide 16, the system correctly identified almost all benign samples, with only one misclassification in that category. The reason we didnt see an increase in false alarms is due to the tuning of our Isolation Forests contamination parameter. By setting it to strictly target extreme statistical outliers, the override acts as a precise priority interrupt rather than a broad filter, allowing us to achieve an overall F1-score of 0.91 and a 91.49% accuracy.
ICACT20260118 Question.2
Questioner: 22-47918-2@student.aiub.edu    2026-02-09 오후 9:36:52 		ICACT20260118 Answer.2
Answer by Auhor sudhakar@ccet.ac.in   2026-02-09 오후 9:36:52
Chrome Click!!
Your hybrid model reached 91% accuracy, which is much higher than the single models. Was this jump in performance because of the 50 selected features or the way the models were combined? Actually, the jump in performance is a result of the synergy between the two, but the Two-Stage Architecture was the primary catalyst. Our hybrid feature selection—using Mutual Information and ANOVA—was essential for stripping away noise from the original 257 features and ensuring we werent dealing with dimensionality issues. This provided a much cleaner foundation for the models to learn from. However, as you can see from our baseline results, even with high-quality features, standalone models like Random Forest and XGBoost stayed under 80%. The significant 13% leap to 91.49% happened specifically because of the Anomaly Override. By adding that second layer with the Isolation Forest, we created a mechanism that specifically catches and corrects edge cases or statistical outliers that typical supervised classifiers struggle to generalize. So, to summarize: the feature selection made the models efficient, but the hybrid combination made the entire system resilient.
ICACT20260118 Question.3
Questioner: arjunvirvashisht@gmail.com    2026-02-10 오전 10:19:42 		ICACT20260118 Answer.3
Answer by Auhor sudhakar@ccet.ac.in   2026-02-10 오전 10:19:42
Chrome Click!!
Do you plan to integrate deep learning models like BiLSTM into your two-stage architecture in the future? In the current work, future extensions are framed around adding dynamic behavioral properties, testing on live malware traffic, and evaluating robustness to adversarial and zero‑day samples, but integrating deep models like BiLSTM into the two-stage RF–XGBoost–Isolation Forest stack is not explicitly listed as planned future work. We plan to extend the two-stage RF–XGBoost–Isolation Forest framework by integrating deep learning models such as BiLSTM to capture sequential behavioral patterns in malware and further improve detection performance.