中文摘要：

基于新浪新闻数据,对热点新闻的连续发表事件时间间隔序列进行了统计分析,以探究新闻内容的选择机制。实证发现该时间间隔分布在个类与总体层面上都遵循带指数截断的幂律分布,由此提出一种考虑时效性的,并基于严格优先及偏好优先选择混合机制的队列模型来揭示新闻选择背后的机制。该模型的数值模拟结果与实证统计数据较好地吻合,表明该模型规则在一定程度上可用于解释新闻报道中出现的非泊松时间特性。

英文摘要：

In this paper, based on the news data of Sina website, inter-event time interval sequences of hot news publication are analyzed to reveal the hidden rules of news selection. Empirical analysis shows that the distributions of the inter-event time intervals between two consecutive news with common keywords follow power-law-like distribution with exponential cutoff both on individual level and aggregated level. Focusing on this finding, we propose a timeliness-based queuing model with mixed mechanisms of strict and preferential priority selections to reveal the hidden principle of news selection. The model results are generally in agreement with the empirical findings, indicating that the proposed model can explain the emergence of non-Poisson properties in news reports.