Feature-Based Semi-Supervised Learning to Detect Malware in Android

Malware is extremely deleterious to an Android operating systems alike desktop operating system. The more Android devices grow, the more we have experienced the growth of Android malware. It does not only pose a serious security threat to user privacy but also lessens the trust

Project Title

Feature-Based Semi-Supervised Learning to Detect Malware in Android

Project Area of Specialization

Artificial Intelligence

Project Summary

Malware is extremely deleterious to an Android operating systems alike desktop

operating system. The more Android devices grow, the more we have experienced the growth of

Android malware. It does not only pose a serious security threat to user privacy but also lessens

the trust on security policies of Android devices. Frameworks and virus protection software can

detect known malware signatures and although, recently, there have been advancements in

detecting unknown malware signatures as well using supervised machine learning approaches,

these methods require huge set of labeled data to train the machine. To solve this problem, we

propose a framework that uses semi-supervised machine learning techniques on API call logs to

detect malware in Android apps. In this approach, feature sub-set selection methods will be

implemented to select suitable features, which will subsequently be utilized to develop an efficient

malware detection model.

Apart from that, statistical validation will also be performed to demonstrate the accuracy of our

model in comparison to supervised machine learning based malware detection models.

Project Objectives

To determine unknown malware in Android apps.

Use minimum human effort to design a model.

To validate the accuracy of your model against supervised learning model that are already developed.

Project Implementation Method

Training Section:

Pre-processing, Mining algorithms, Association rules.

Detection Section:

Permission scanning, Filtering, Rules Matching( association rules)

Results:

Normal and abnormal.

Benefits of the Project

Detect unknown attack accurately using less computational resources effectively detect unknown malware.

Effectively detect unknown malware.

Quickly detection of malware.

Obfuscation and encryption technique will not effect detection.

Faster feedback to the user .

Reducing the storage space usage.

Reduced the complex processing.

Less complex to analyse.

Technical Details of Final Deliverable

Android Emulator, Android Apps, Machine Learning Algorithms, API calls

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Security

Core Technology

Big Data

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Responsible Consumption and Production

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com