<![CDATA[This study proposes a CNN-LSTM time series analysis-based intrusion detection system (IDS) model to improve accuracy and efficiency in cloud computing environments. With more organizations moving to the cloud, security threats are becoming more sophisticated, rendering traditional detection methods inadequate. The objective of this study is to develop and evaluate a hybrid model that can address these challenges. The methodology used involves an experimental quantitative approach on a representative CSE-CIC-IDS2018 dataset. This dataset underwent rigorous data preprocessing, including data cleaning, conversion to time series format, and feature selection using stationarity and Granger causality tests. The CNN-LSTM model was then trained and evaluated using accuracy and computational efficiency metrics. The results showed superior model performance with an accuracy of 0.910, precision of 0.874, and F1-Score of 0.882. The model also demonstrated good computational efficiency, with a training time of 3.9887 seconds and a prediction time of 0.3607 seconds, making it suitable for real-time detection. This study concludes that the CNN-LSTM model is a viable solution for improving cloud computing security, offering a balance between high accuracy and good computational efficiency. Future research could explore multi-dataset validation and the integration of interpretation methods to improve its application.]]>
Copyrights © 2025
