Proactive News Article Summarization Service Using Personal Intention Models

Abstract

Nowadays, we live with a huge amount of data. By IDC report, the amount of data generated in 2011 is about 1.8 ZB (trillion GB). Ironically, there are small amount of useful information when you are looking at a number of papers, internet articles, movies, pictures, and social network posts. As a result, it is required that users extract useful information manually. However, this is tedious and hard task. To solve this problem, a lot of sophisticated techniques have been proposed to provide summarization service. Although each user has different desire on the summarization, the service uses much computational resource to produce standard summarization for all users. In this study, we propose to use machine learning to predict the intention of users on the summarization. It can reduce the computational cost to summarize all the documents and make available proactive summarization service. To validate our proposal, we run experiments with eight participants for two news portals on five topic areas.