A Novel Framework based on Extra Tree Regression Classifier and Grid Search LSTM for Intrusion Detection in IoT and Cloud Environment
International journal of intelligent engineering and systems · 2024
Currently, the Cloud Computing (CC) and Internet of Things (IoT) have emerged as advanced technologies that enable new levels of connection and data processing.As the IoT ecosystem grows, it becomes more important to ensure the security and integrity of IoT devices and the data they create in cloud.The identification and prevention of intrusions in both cloud and IoT cloud systems has become a major challenge.In this research work, a new intrusion detection framework based on Extra Tree Regression Classifier and Grid Search Optimized Long ShortTerm Memory (ETR-GSO-LSTM) is used to identify and classify intrusions in IoT and Cloud environments.The input data is first gathered via the CIC-IDS-2018 and KDD-Dataset, which include a lot of information about network traffic and possible security issues.The data preprocessing tasks such as label encoding and data augmentation was performed to perform more amount of labelled data for analysis.Another crucial phase in the intrusion detection process is feature selection, and the ETR Classifier has shown to be a significant tool in determining the most relevant characteristics from the dataset.These chosen features assist in reducing dimensionality and improving the accuracy of the intrusion detection model.Finally, GSO-LSTM appears as a potential strategy for classification, which employs the capacity of LSTM networks to examine sequential data and find abnormalities in real time.The proposed ETR-GSO-LSTM achieves detection or classification accuracy of 99.95% and 99.9% on the CIC-IDS 2018 and NSL-KDD datasets, respectively.From the result analysis, it clearly shows that the proposed ETR-GSO-LSTM obtains better performance in all the metrices when compared to Enhanced Long-Short Term Memory with Recurrent Neural Network (ELSTM-RNN) and Unsupervised Technique Ensemble based IDS (UTENIDS).