<p>With the advancement of technology worldwide, the production of vast amounts of data has led to the rise of cloud computing. This evolution has improved the reliability of applications and has helped address problems that arise from the continuous influx of data. However, as more people have started using cloud services, issues such as latency and security have emerged. Experienced researchers have been working to address these challenges. This study focuses on improving security in the cloud platform, particularly against distributed denial of service attacks, which are among the most common threats and can cause significant damage. Machine learning and deep learning techniques were utilized to build an intrusion detection system capable of detecting these attacks in real time with high accuracy. The accuracy of this model was evaluated using various feature selection techniques, including random forest (RF), analysis of variance (ANOVA), permutation, mutual information, and a hybrid model. Seven machine learning models were used for comparison: decision tree, random forest, XGBoost (extreme gradient boosting), logistic regression (LR), categorical boosting (CatBoost), and gradient boosting (GBoost). The results indicated that most of the models achieving accuracy of up to 92% and a good precision rate while significantly reducing false positives.</p>