Malware Detection Inside App Stores Based on Lifespan Measurements

Potentially Harmful Apps (PHAs), like any other type of malware, are a problem in the modern Android ecosystem. Even though Google tries to maintain clean app ecosystem, Play Store is still one main vectors for spreading PHAs. In this paper, we propose solution based on machine learning algorithms detect PHAs inside application markets. Being markets entry vectors, capable detecting submitted or submission those needed. This an market and can be used as filtering method, automatically block publishing novel The proposed static analysis, even several analysis solutions have been developed, innovation system its training creation dataset. We created new dataset that uses criteria lifespan Play, shorter time active higher probability PHA. criterion was added order avoid usage bias antivirus engines malware. Involving method detection does not replicate existing engines. Experimental results proved obtains 90% accuracy score, using 91,203 applications published Store. Despite showing decrease accuracy, compared with models focused PHAs; it necessary take into account complementary different method. presented work combined dynamic models, since drastically different, measurements.