User’s Age Detection Using Smartphone Motion Sensor Data

At the age of two, Children are smart enough to be able to use Smart devices. According to studies, children who spend more time on phones are exposed to health issues; moreover, accessing unwanted content leads to social and behavioral changes. To automatically regulate access to unwanted content on the internet, it is required to identify the age group of the user. In this work, the age of smartphone users is detected based on their implicit behavior in the form of the motion sensor reading. The dataset used is a publicly available dataset that consists of 100 child users and 100 adult users. Different Machine learning models such as KNN, LR, SVC, and Voting have been attempted to detect user age based on motion sensor readings. KNN offers the best performance with 97.30% F1-Score and 98.86% accuracy.