News Popularity Prediction with Local-Global Long-Short-Term Embedding.

Shuai Fan,Chen Lin,Hui Li,Quan Zou

DOI: https://doi.org/10.1007/978-3-030-91560-5_6

2021-01-01

Abstract:Predicting news popularity is an essential topic in the news industry. It is challenging because numerous factors influence public response to the news. This paper presents F-4, a neural model to predict news popularity by learning news embedding from global, local, long-term and short-term factors. F-4 integrates a sentence encoding module to represent the local context of each news story; a heterogeneous graph-based module to capture the short-term information propagation from current buzz words to each news story; a sequential module to extract long-term popularity features in entity sequence; and an attention module to learn global news-entity correlations. Extensive experiments on real-world Chinese and English news datasets demonstrated that F-4 outperforms state-of-the-art baselines in predicting and ranking news popularity.