A Combined Leading Ensemble Decision Classifier Module (CLEDCM) for Intrusion Detection in IoT

The Internet of Things (IoT) suffers from different types of attacks due to the vulnerability present in devices. Cyber security has become a critical concern in protecting IoT networks from malicious activities and ensuring the privacy and integrity of the data transmitted. It also develops a demand for advanced technology procedures to boost the security of IoT networks. One approach gaining attraction in addressing this challenge is by deploying ensemble machine or Deep learning models for identifying the attack categories and then preventing them, in cyber security. In the Ensemble mechanism, every model or framework of the Machine learning classifier will be trained on multiple subsets of features and data instances within controlled ratios, making them capable of capturing different aspects of the attack patterns. In ensemble techniques the best-predicted values of every individual classifier will be merged up to make up a superior model with great accuracy level so that the predictions ever made will match up to great accuracy and model needs as well. This study combines the results and metrics fetched from 4 base classifiers LightGBM, Random Forest classifier (RFC), XGBoost, and CatBoost, and merged it into a Combined Leading Ensemble Decision Classifier Module (CLEDCM) for the prediction of every record or tuple to find the accurate class it belongs to in case of Multiclass classification scenarios on IOT based Datasets (i.e. NSL-KDD, NF-BOT-IOT datasets). CLEDCM merges the prediction from all the major 4 classifiers in such a way, that the overall Accuracy % and f1-score will get increased by fetching the best prediction from all base classifiers in a group individually Class-wise. In the results, we observed that our proposed model got the highest accuracy of 99.94% for NSL-KDD test data for Multiclass classification in comparison to other base classifiers.