Authors:- Shilpa Mahajan, Sakashi Tanwar
Abstract:- It is incredibly practical to reduce the amount of work required to battle malware. We present a flexible, semi-automated approach for analysing large datasets of Android apps and identifying novel malware families. Until 2010, the industry standard for pest detection was the apps rely heavily on signatures. Because every minor modification in malware renders it useless, new signatures are frequently produced, a procedure that necessitates a significant amount of time and resources from knowledgeable professionals. With the architecture we present, applications may be automatically classified into families and formal rules may be proposed to identify them with 100% recall and reasonable accuracy. The families are used to either securely broaden the experience of specialists on new samples or to limit the number of applications that require extensive analysis. We demonstrated the efficacy and scalability of the existing technique. Experiment with 1.5 million Android apps in a database. In 2018, the structure was successfully sent to Koodous, a community-oriented malware stage.
Keywords:- Android, malware, antivirus, application
Full Issue
| View or download the full issue | PDF 77-82 |