Automatic Thai news summarization using principle of summarization and interpretation template

Abstract

Online news is a well known application of information delivery in the age of digital society. There is usually time consumption to explore all sources of the online news to consume the information which is usually written with different writing styles to report the same event. Especially, in the case of Thai online news, synonyms, acronyms, superfluous words, and jargon words generally cause a redundancy roblem for automatic text summarization. To deal with this problem, this thesis proposes a method for summarization of the Thai online news using a principle of summarization and an interpretation template. According to the principle of the summarization, the interpretation template is generated and filled with the common content extracted from different sources of the online news. To evaluate the quality of the summary, the evaluation task is set up to assess the satisfaction of the readers by five evaluators. The evaluation results show that the proposed method obtains 80%, 73%, 73% and 73 % of grammar, conciseness, clarity and coherence scores.