스토리텔링 저작을 위한 서사정보 추출 시스템 설계 및 구현

Abstract

Recently, the perceived value of stories has risen, which has created a greater focus on storytelling. Storytelling is method used for immersing an audience and prompting interest in the development of plot. This method has been receiving a lot of attention recently as an important technique in the cultural industry.

More specifically, there have many more cases of stories being created using authoring tool systems through Human-Computer Interaction (HCI). Compared to other countries, there have not been many studies regarding using such systems for storytelling in Korea.

This thesis proposes a system which automatically extracts narrative information, including key storytelling elements, that comprises the main components of a written story. Patterns for extracting narrative information nouns within the text are defined by applying a Korean natural language process technique to story data collection and analysis. Story data is then extracted using term frequency(TF).

To assess this narrative information extraction system, experiments were done to analyze data from six stories. The results determined using the automated character extraction method. The results showed that approximately 71.83% of the average precision ratio and performance had been confirmed.