With the increasing complexity of the cyber-attacks, traditional intrusion systems have become inadequate in effectively detecting and preventing network attacks. Modern machine learning techniques have the capability of preventing sophisticated and aggressive cyber-attacks and have emerged as a promising approach to enhance the accuracy and efficiency of the intrusion detection. This research paper provides a comprehensive analysis and evaluation of the integration of machine learning and deep learning algorithms into intrusion detection system with respect to the Distributed Denial of Service (DDoS) attack. The study begins by presenting an overview of intrusion detection systems, highlighting their importance and significance in safeguarding network infrastructure. Subsequently, it explores supervised machine learning algorithms and deep learning models, such as Random Forest tree, Support Vector Machines (SVM), ResNet and TabNet. The paper presents the pros and cons of existing IDS models with respect to real-time processing, scalability, false positive and false negative results, and IDS vulnerabilities. The above-mentioned algorithms have been trained and evaluated with CICIDS-2017 dataset. The performance metrics used to evaluate the IDS models are accuracy, precision, F1 score, training time of each model, and adaptability to evolving attack patterns.