• ENSEMBLE-ENHANCED THREAT INTELLIGENCE NETWORK (EETIN): A UNIFIED APPROACH FOR IOT ATTACK DETECTION

Ajay Chandra MK

Abstract:
The procedure of identifying and reacting to unauthorized or malicious behaviours within the IOT system, is called IOT attack detection. There are some several advantages and challenges associated with implementing and maintaining the detection systems. Some of the limitations are Complexity, false positive and negative and scalability. To defeat these restrictions, we suggested a deep-learning based classification model (INTE). The proposed methodology consists of (1) Pre-Processing (2) Customized Dimensionality Reduction (3) Feature Extraction (4) Deep Learning-based Classification via Intelligent Network Threat Ensemble (INTE) Model (5) Evaluating and Testing. In pre-processing, the gathered data are pre-processed via data cleaning and Transformation. The pre-processed data are used to dimensionality reduction using O-PCA. The principal components are optimized via TLBO (Teaching learning-based optimization). Then the dimensionality reduced data are utilized to filter the relevant features using Time and frequency domain features and non-linear features. The extracted features are used to classify the attack detection via INTE model which is a combination of CNN, RNN and transformer-based model. And, the model will be evaluated using the performance metrics. The proposed methodology is executed using MATLAB.