Adaptive Student Interest Prediction in Mobile Learning Apps Using Recurrent Neural Networks

The education field uses portable software systems as mobile applications that operate on smartphones and tablets to deliver interactive, accessible services across different domains. Mobile learning applications now enhance student engagement by delivering dynamic content alongside real-time assessments within learning environments tailored to user needs. This study presents a combined Machine Learning (ML) solution to address the limitations of student interest prediction systems. The proposed approach begins with data preparation, utilizing Principal Component Analysis (PCA) to reduce dimensionality while extracting meaningful features from the student activity log. The analysis employs Support Vector Machines (SVM) to create student interest classes by evaluating their behavioral patterns. The system utilizes Recurrent Neural Network (RNN) technology to analyze student interaction timelines and create customized teacher or content suggestions. The method presents an efficient system that supplies precise time-oriented interest observations for improved adaptive learning assistance in mobile frameworks.