kumoh national institute of technology
Networked Systems Lab.

Fabliha Bushra Islam, Rubina Akter, Jae-Min Lee, and Dong-Seong Kim, "Deep Learning Based Network Intrusion Detection for Industrial Internet of Things", 2020 Korean Institute of Communication and Sciences (KICS) Summer Conference, August 12-14, 2020, Yong Pyong Resort, Pyeongchang, Gangwon Province, Korea, (N8)
By : Bushra
Date : 2020-06-07
Views : 142



This paper represents a deep learning model dependent on anomaly detection techniques for malicious discovery in network traffic configuration assembled from TCP/IP packets. Four Artificial Neural Network (ANN) model variations were compared using a public dataset called NSLKDD. The ANN was implemented using R-programming, where in, input layer, hidden layer and output layer were manipulated to select the best model with optimal performance in terms of reduced Mean Absolute Percentage Error (MAPE). The experimental results illustrate that model 1 outperformed other ANN models with an accuracy of 96.74%. The conclusion of this paper is that ANN model offers a good performance for malicious detection. However, caution should be on design of ANN models as we observed that layer numbers have impact on their accuracy. This proposed ANN model can be effective for Industrial IoT networks, and smart factories management to restrain future malicious attacks.