Analyzing the Narrative Structure in a Fantasy Fiction Using Text Mining Technology

• Main Authors: Wei-Liang Liu, 劉偉樑
• Other Authors: Sung-Chien Lin
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/90539069283823486726

碩士 === 世新大學 === 資訊傳播學研究所(含碩專班) === 103 === Fantasy Fiction is a type of fiction that full of unlimited imagination, which stories are told with myths or legends. Many fantasy fiction works had been adapted for types suitable to movies, TV series or video games in the recent years. However, a fantasy fiction usually comes with millions of word and the story in it is so complicated, such that it is difficult to be understood in a glance and to be adapted easily. This study proposed a method that helps solve the difficulty of analyzing a fantasy fiction by using a series of Natural Language Processing technique to explore the narrative structure inside the fiction. Major algorithms of the proposed method include the Maximum Matching Algorithm for Chinese Word Segmentation and the Latent Dirichlet Allocation Algorithm to extract the topics inside the text of the fiction. The text of the Chinese version of Dragonlance Trilogy was used as the data for testing the method. This study also analyzed the relation between the resulting topics and the original chapters of the fiction to present its narrative context.