Appformer: A Novel Framework for Mobile App Usage Prediction Leveraging Progressive Multi-Modal... (
Original paper: https://arxiv.org/abs/2407.19414
Summary of ArXiv paper 2407.19414:
The paper introduces Appformer, a novel framework for mobile app usage prediction that leverages progressive multi-modal data fusion and feature extraction. The authors, Chuike Sun, Junzhou Chen, Yue Zhao, Hao Han, Ruihai Jing, Guang Tan, and Di Wu, propose this approach to address the challenges of accurately predicting future app usage patterns based on historical data. The Appformer framework is designed to integrate various types of data, including user behavior, app characteristics, and environmental factors, which are fused progressively to capture complex relationships between these modalities. This fusion process involves multiple stages, each aimed at extracting meaningful features from the combined data sets. The authors highlight the importance of considering both short-term and long-term dependencies in app usage patterns, as well as the impact of external factors such as time of day, location, and device type. They also emphasize the need for a flexible and adaptive prediction model that can accommodate changing user p