Improved Machine Learning Algorithm for Detecting Intrusions in Web-based Applications
Background: Intrusion is defined as an unauthorized access to sensitive information or data. When this happens, security is breached and confidentiality of the affected system is in doubt. With the proliferation of web applications, intrusion detection systems have increasingly become relevant in today's digital age. Aim: Existing intrusion detection systems have low accuracy and precision rates which causes false alarm. Therefore, the aim of this research is to develop an improved algorithm that has the capacity to correctly detect intruders at the right time Method: Real-world datasets collected from a forensic framework was used to evaluate the performance of the proposed algorithm. The dataset was preprocessed to remove redundant or irrelevant features which were normalized. The datasets were then divided into training and testing sets. Results: The proposed algorithm outperformed existing ones with accuracy of 98.89%. Research findings suggest that, the proposed algorithm is effective in detecting intrusions in web-based applications
Download Paper