Please use this identifier to cite or link to this item: https://bspace.buid.ac.ae/handle/1234/1586
Title: A Digital DNA Sequencing Engine for Ransomware Analysis using a Machine Learning Network
Authors: KHAN, FIROZ
Keywords: ransomware
DNA sequence
machine learning
malware
Grey Wolf Optimisation
binary search
machine learning network
Issue Date: Feb-2020
Publisher: The British University in Dubai (BUiD)
Abstract: The research work proposes a novel detection mechanism for ransomware using machine learning approach using Digital DNA sequencing. The proposed work contains three significant phases: Preprocessing and Feature Selection, DNA Sequence Generation and Ransomware Detection. In the first phase, data preprocessing and feature selection technique is applied to the collected dataset. The preprocessing of data includes remove missing value records and remove columns that have a negligible impact. The feature selection uses Grey Wolf Optimisation and Binary Search algorithms for choosing the best features out of the dataset. In the DNA Sequence generation phase uses design constraints of DNA sequence and k-mer frequency vector. A newly collected dataset after feature selection is used to generate the DNA sequence. In the final phase, the new dataset is trained using active learning concept, and the test data is generated using a random DNA sequence method. The data is finally classified as either ransomware or goodware using the learning methodologies. The results are found to be promising and reconfirm the fact that the developed method has efficiently detected ransomware when compared to other methods. The thesis concludes by a discussion of future work to advance the proposed method and future directions of research on the use of Digital DNA sequencing engine for general malware detection.
URI: https://bspace.buid.ac.ae/handle/1234/1586
Appears in Collections:Theses for Computer Science

Files in This Item:
File Description SizeFormat 
2016146087.pdf3.67 MBAdobe PDFThumbnail
View/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.