Characterization and Early Detection of Evergreen News Articles.

摘要: Although the majority of news articles are only viewed for days or weeks, there a small fraction that read across years, thus named as evergreen articles. Because maintain timeless quality and consistently interests to public, understanding their characteristics better has huge implications outlets platforms yet few studies have explicitly investigated on Addressing this gap, in paper, we first propose flexible parameterized definition capture long-term high traffic patterns. Using real dataset from Washington Post, then, unearth several distinctive build an early prediction model with encouraging results. less than \(1\%\) were identified evergreen, our achieves 0.961 ROC AUC 0.172 PR 10-fold cross validation.