Hot Events Analysis based on Japanese Corpora Processing
Download as PDF
DOI: 10.25236/icsemc.2017.02
Author(s)
Wang Tong, Yi Mianzhu
Corresponding Author
Wang Tong
Abstract
Lacking of study on sufficient foreign literature and data, domestic researches on hot events between China and Japan are one-sided due to the gap between languages. In order to analysis hot events more objectively by Japanese corpus processing and knowledge mining, this paper puts forward a method of name recognition and reference resolution based on surname dictionary and syntax rules. What is more, a characters detection model of hot events is designed and the knowledge graph of core characters is showed on map database. Finally, we tested the Japanese name recognition method on Asahi Shimbun corpus, of which the accuracy was 90.07% and the recall rate was 91.15%.
Keywords
Japanese name recognition, Map database, Hot events analysis, Asahi Shimbun, Yasukuni Shrine.