| Autor: |
Pitty, Nagarjuna, Praveen, R. V. S., Jain, Virendra, Tamilselvam, M., Haripriya, D., Bansal, Saloni |
| Zdroj: |
Library of Progress-Library Science, Information Technology & Computer; Jul-Dec2024, Vol. 44 Issue 3, p5644-5663, 10p |
| Abstrakt: |
Popular IoT devices such as smartphones have dramatically enhance the network connection that has posed new unique security threats that calls for more developed security systems to protect connected systems. Thus, this research aims at comparing the performance of different machine learning algorithms that are K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Random Forest (RF), and Deep Neural Networks (DNN) to address the anomalies identification and IoT security improvement. These algorithms were used in the study to assess the IoT traffic and attack data set. This is due to the fact that results which were obtained proved that DNNs had the highest level of accuracy of about 97. 5%, RF with 93, Breast with 90 and GYN with 87. 2%, SVM with 89. 7% and K-Nearest Neighbors with 85%. 4%. While the DNN had a better accuracy than the other models, the model was more computationally intensive compared to the other one; RF is a good trade-off between accuracy and time. KNN while being computationally cheap had the lowest accuracy and higher FPR. Comparison of the proposed work with already published literature also validates that as the present day algorithms provide rudimentary level of security, this study reveals that enhancing and combining these techniques is essential to enhance real-time detection and robustness of the systems. This present research aims at adding to the existing literature on the subject of effective protection of IoT systems with adequate cybersecurity measures. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
