Automatic prediction method for hot news topics based on time-series analysis

作者：Gu, Wanrong; Dong, Shoubin^*; He, Jinchao

来源：Journal of Computational Information Systems, 2014, 10(8): 3473-3485.

DOI：10.12733/jcis9972

• Automatic prediction
• Exponential smoothing predictions
• News information
• News topics
• News websites
• Prediction model
• State-of-the-art approach
• User's preferences

Hot news topic is one of the important characteristics in Internet news websites nowadays. News topics with high quality and focus can enhance experience and save uses'time. Most of existing methods rely on artificial or statistical rate of users'click. However, user's preferences are often changing over time, especially the demand of news information. In this paper, we propose a time-series analysis based hot news topics prediction model, which is a combinative model for hot news topic prediction using Triple Exponential Smoothing Prediction Model (TESPM) and Prediction Model Based on Spectral Analysis (PMSA) methods. The experimental results show that our model outperforms good performance for hot news topic prediction compared to the state-of-the-art approaches.

• 出版日期2014-4-15
• 单位华南理工大学