Comparison of detection model using machine learning on android malware / Muhammad Ikmal Ihsan
Mobile devices have experienced tremendous growth during the past ten years. As gadgets become more pervasive and people save more sensitive data on their mobile devices, the prevalence of mobile malware has increased. Malicious software, commonly known as malware, poses a greater risk to these mobi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Student Project |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/83340/2/83340.pdf https://ir.uitm.edu.my/id/eprint/83340/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Teknologi Mara |
| Language: | English |
| Summary: | Mobile devices have experienced tremendous growth during the past ten years. As gadgets become more pervasive and people save more sensitive data on their mobile devices, the prevalence of mobile malware has increased. Malicious software, commonly known as malware, poses a greater risk to these mobile devices nowadays. Recently, several articles have been published regarding the proliferation of Android malware. Many modern technologies, such as smartphones, have been used into Android malware development, enabling it to advance. It has been used for a long time but is now worthless due of the evolution of Android malware and the inability to detect it. This project utilized supervised machine learning techniques such as SVMs, Naive Bayes and Random Forest to build an android malware detection model. It also evaluated and train the selection of Android characteristics to evaluate the malware detection model's performance. Then, the project examined the effectiveness of several machine learning detection models in identifying Android malware. This project has been divided into five distinct parts, each with a distinct purpose. Initialization, planning, development, evaluation, and documentation are all part of the process. In the end of the project, the result has been discussed and been compared for each machine learning used to get the highest accuracy to achieve the project objectives. The result of the comparison using the machine learning techniques for Android malware dataset discovered that SVM machine learning get the highest percentage with accuracy of 0.93. It also recorded that SVM machine learning got the lowest FPR and highest TPR among other machine learning used in the project. For future references to do this project, the project can be improved by using the project’s own Android malware dataset and use more than three machine learning when training and evaluating the dataset to discover the true potential for each machine learning |
|---|