Semi-Supervised Machine Learning Approach for DDOS Detection

Semi-Supervised Machine Learning Approach for DDOS Detection

2111CS020341 - D.POOJITH KUMAR                          2111CS020342 - B.POOJITHA

2111CS020343 - P.PRADEEP                                        2111CS020344 - M.PRADEEPTHI

2111CS020345 - A.PRANAVI

Guided by

Prof. Mohammad Abdul Najeeb

1.Abstract

            The proliferation of malicious applications poses a significant threat to the Android platform, as numerous applications exploit network interfaces to surreptitiously obtain users’ personal information and initiate malicious attacks. This paper presents an effective and automated methodology for detecting malware by analyzing network traffic employing text-based approaches. Specifically, we characterize each HTTP flow originating from mobile applications as a text document and apply Natural Language Processing (NLP) techniques to extract pertinent textual features. We employ the N gram method from NLP to analyze traffic flow headers and apply a chisquare test for feature selection to identify the most pertinent features for malware detection. This approach elucidates the correlation between traffic patterns and malicious behavior. Our solution presents a novel methodology for detecting malware by treating mobile network traffic as text documents and applying text analysis methods for reliable detection.