Mostofa Md Golam

Passionate about the intersection of education, technology, and research, I am a dedicated professional with a keen interest in advancing learning through innovative teaching methods and cutting-edge technologies. My professional journey is driven by a passion for exploration, education, and the transformative potential of technology in shaping the future of learning.

• Yan Zhang, Lecturer, Information Technology, Charles Darwin University, 08 8946 6958, yan.zhang@cdu.ede.au
• Sami Azam, Professor, Information Technology, Charles Darwin University, 08 8946 7646, Sami.Azam@cdu.edu.au

• Data Warehouse, https://lnkd.in/gzH6dzwd
• Microsoft Office (Word, PowerPoint, Excel)
• Python for Beginners, https://drive.google.com/file/d/18C3gMfKnOa-h6Tkv77IjzGxQh7xyYUTo/view?usp=drive_link
• Django for Beginners, https://drive.google.com/file/d/10xyNjNF_j0imI_ITvTbVn6NunM-p9Ktw/view?usp=drive_link
• Cyber Security, https://drive.google.com/file/d/1CcMZAtDfs_qC3uMPLoQKWkHYgb1JwgL2/view

Ensemble Feature Selection for Android SMS Malware Detection, International Conference on Cyber security, Cybercrimes, and Smart Emerging Technologies, 05/10/22 to 05/11/22, Riyadh, Saudi Arabia, Our research paper was published on 'International Conference on Cyber security, Cybercrimes, and Smart Emerging Technologies'. In this thesis, we proposed an ensemble feature selection method for Android malware detection. We used six feature selection methods such as information gain, chi-square, RFE, ANOVA-F, forward selection, and backward elimination to select features. Then, the most common 12 features among all the feature selection methods are selected. Five classifiers, named decision tree, random forest, gradient boosting, XGB, and KNN, are trained against the selected 12 features. It was found that the gradient boosting based classifiers outperformed the other classifiers. While the XGB outperformed other classifiers in ransomware and adware classification, the gradient boosting method outperformed all other methods in the detection of SMS malware and scareware. The proposed method achieved up to 22.17% higher performance gain compared to the state-of-the-art methods. The number of features is reduced by 33% to 52% depending on the type of malware., https://link.springer.com/chapter/10.1007/978-3-031-21101-0_2