Abstract:Current state-of-the-art Online Video Services (OVSs) need to simultaneously serve hundreds of millions of users at peak time. To improve service efficiency, the caching strategies of their Content Delivery Networks (CDNs) usually work with video segments of a couple of seconds in length, which poses great challenges in popular prediction of video segments. Popularity prediction of entire videos is generally based on the correlation of early views and future views. However, for the popularity prediction of video segments, the first and the rest segments of a video should be distinguished first because users usually start watching a video from its first segment, but could skip any of the rest ones. Towards this end, we propose a novel method for video segment popularity prediction, in which, the popularity of videos' first and rest segments are predicted with different models. For the first segments, the popularity is predicted like entire videos by using a Multi-Linear Regression (MLR) Model based on past popularity. And for the rest segments, as the jumping viewing behavior has largely weaken the indicating power of their early views on future views, we predict their popularity based on the recent viewed previous segments. Specially, considering the real-time requirement of CDNs caching, it is infeasible to investigate the viewed segments sequences of users with sequence models, hence the prediction model is built on statistic level. Moreover, to maintain the performance in online prediction, we equip the two prediction models with online learning capability so as to timely capture the changes in popularity. For promptly service, the job is fulfilled with an online compensator instead of retraining the two prediction models with revealed target of past predictions. The online compensator estimates the prediction errors of the two models through a LSTM network based on their recent errors and the popularity data corresponded to recent tasks. Through extensive experiments on real-world data of iQIYI, one of the most popular OVS providers in China, we have demonstrated the promise of the proposed method.

Program Popularity Prediction Approach for Internet TV Based on Trend Detecting.

Big Data Analytics for Program Popularity Prediction in Broadcast TV Industries.

Program Popularity Prediction Model of Internet TV Based on Viewing Behavior

Time Context-Aware IPTV Program Recommendation Based on Tensor Learning.

Research on efficient organization and browsing of TV program based on new media platform

Predicting the Popularity of Viral Topics Based on Time Series Forecasting

Hybrid Machine Learning Approach to Popularity Prediction of Newly Released Contents for Online Video Streaming Service

A Real-Time Method to Predict Social Media Popularity

Prediction of Online Topics’ Popularity Patterns

Sequential recommendations based on time series and dynamic interest for TV program

Online Video Popularity Regression Prediction Model with Multichannel Dynamic Scheduling Based on User Behavior

A Novel Time Series Approach for Predicting the Long-Term Popularity of Online Videos

Lifetime popularity prediction for online videos

Popularity prediction in microblog based on LR-DT

Predicting the Online Performance of Video Service Providers on the Internet

TV shows popularity prediction of genre-independent TV series through machine learning-based approaches

Modeling Periodic Impulsive Effects on Online TV Series Diffusion

Online Popularity Prediction of Video Segments Towards More Efficient Content Delivery Networks

EPAB: Early Pattern Aware Bayesian Model for Social Content Popularity Prediction

Cost-Effective User Monitoring for Popularity Prediction of Online User-Generated Content

Predicting Key Events in the Popularity Evolution of Online Information