Implicit authentication method for smartphone users based on rank aggregation and random forest

Currently, the smartphone devices have become an essential part of our daily activities. Smartphone’ users run various applications (such as banking and e-health Apps), which contains very confidential information (e.g., credit card number its PIN). Typically, smartphone’s user authentication is achieved using mechanisms (password or security pattern) to verify identity. Although these are cheap, simple, quick enough for frequent logins, they vulnerable attacks such shoulder surfing smudge attack. This problem could be addressed by authenticating their behaviour (i.e., touch behaviour) while smartphones. Such behaviours include finger’s pressure, size, pressure time tapping keys. Selecting features (from behaviours) play important role in process’s performance. paper aims propose efficient method providing implicit not imposing additional cost special hardware addressing limited capabilities. We first investigated feature selection techniques from filter wrapper approaches then used best one method. The random forest classifier evaluate techniques. It also achieve classification task Using a public dataset, experimental results showed that filter-based technique rank aggregation) build environment. accuracy around 97.80% only 25 out 53 require less mobile resources (memory processing power) authenticate users. At same time, has error rate: 2.03 FAR, 0.04 FRR, 1.04 ERR, comparing related work. These promising would develop application allows legitimate owners avoiding traditional problems fewer resources.