← Back to Publications List

A Multimodal Prediction Model for Smartphone Addiction Among Children Using Continuous Authentication Through Behavioral Biometrics

Students & Supervisors

Student Authors
Sudeep Mondal Deep
Bachelor of Science in Computer Science & Engineering, FST
Supervisors
Aneem Al Ahsan Rupai
Assistant Professor, Faculty, FST
Sazzad Hossain
Assistant Professor, Faculty, FST
Md. Reazul Islam
Assistant Professor, Faculty, FST

Abstract

The problem of smartphone addiction in children has become increasingly critical, and signing in to the device excessively leads to emotional, cognitive and behavioral problems. Conventional question and answer based instruments like the Smartphone Addiction Scale (SAS) are helpful, yet they are intensive to opinions and do not reflect dynamic patterns of interaction. This paper suggests a sensor-driven system based on the principles of Continuous Authentication (CA) to monitor the levels of smartphone addiction based on the natural touch and motion patterns of children. An application created in Android took into account touchscreen, acceleration, and gyroscopes data of children and then underwent formal preprocessing and statistical feature extraction. Three machine learning models including XGBoost, random forest and SVM were trained and assessed using the same features. As experimental outcomes show, XGBoost performed the best with an accuracy of 95%, an AUC of 0.99, and a misclassification rate of 83 true and 88 true. On the whole, the results show that behavioral biometrics may be used as an objective and a continuous method of early identification of smartphone addiction in children which can be utilized by parents, educators and researchers.

Keywords

Smartphone Addiction Children Continuous Authentication Behavioral Biometrics Touch Gesture Analysis Hand-Waving Patterns Machine Learning Multi-modal Authentication Smartphone Sensors

Publication Details

  • Type of Publication:
  • Conference Name: 5th International Conference on Mathematical Modeling, Computational Intelligence Techniques and Renewable Energy (MMCITRE-2025)
  • Date of Conference: 10/12/2025 - 10/12/2025
  • Venue: IIIT Naya Raipur, Chhattisgarh, India
  • Organizer: International Institute of Information Technology, Naya Raipur (IIIT Naya Raipur)
View on Publisher