Deep Learning and Machine Learning-Based Approaches to Inferring Social Media Network Users' Interests from a Missing Data Issues

This research explores how user-provided information can be leveraged to identify evolving user interests and generate tailored recommendations. It addresses the challenge posed by inactive users who rarely share or interact on social media, making it difficult to assess their profiles due to insufficient information. To infer and forecast the interests of these inactive users, the study examines how individuals' interests can be deduced from their friends' activities on social networks. The proposed method utilizes machine learning and deep learning to analyze data from user friends and predict interests, even when information is missing. This approach is evaluated using the standard Kaggle Social Network dataset and performance metrics such as recall, accuracy, precision, and F-measure (F1). The findings demonstrate the effectiveness of the proposed user detection strategy.