Statistical Analysis of Big Data Models in Android Malware Detection

The term "malicious software" (or "Malware") refers to any software designed specifically to cause harm to a computer system. Malware has also progressed in its ability to avoid detection and launch attacks in tandem with the development of computing technology. Attackers are motivated to target Android vulnerabilities due to the growing popularity and reliance on smartphones. There are several types of stealth software, but none are as dangerous as ransomware. Ransomware is a type of malicious software designed to hold users' files and systems hostage in exchange for payment. Numerous ransomware assaults may have been inspired by the meteoric rise in mobile commerce. Desktop PCs and mobile devices are both vulnerable to malware. Microsoft Windows is the most used OS for home desktop computers across the world. As a result, a breach in one system might potentially affect many others. This research study offers a multi-stage system in which the initial stage makes use of features that are simpler to extract with cheaper classifiers. Level 2 is considered for detection if the confidence of the level 1 classifier is below a predetermined threshold. The proposed study is not only concerned on accuracy but also on cost, because integrating many classifiers may not be possible due to the additional memory and computing requirements. In addition, the time to output is crucial for the practicality of a malware detection tool. So as to strike a balance between accuracy and cost, multilevel ensemble system with a modest number of classifiers has been considered.