The field of quality of experience (QoE) and service predictions is moving towards more accurate and efficient models, leveraging machine learning and tensor factorization techniques. Researchers are exploring alternative methods to traditional quality of service (QoS) metrics, focusing on dynamic and real-time predictions. Notably, the use of lightweight generative modeling frameworks and non-negative tensor factorization is gaining attention. Innovative approaches are being proposed to capture temporal patterns, interest diffusion, and user behavior in various applications, including video streaming and social activity data streams. Some noteworthy papers in this area include:

• Generative QoE Modeling, which introduces a lightweight approach using Vector Quantization and Hidden Markov Models for temporal sequence modeling.
• D-Tracker, a method for continuously capturing time-varying temporal patterns within social activity tensor data streams and forecasting future activities, showcasing high forecasting accuracy and scalability.