A cinemetric approach to sentimental processing on story-oriented contents

Most of stories are usually bsed on many kinds of relationships among the characters. particularly, in various digital contents, to efficiently manage, we present a novel cinemetric approach to exploit a social network (called Character-net) extracted from the stories. Since story transcripts are composed of several elements (e.g., scene headings, character names, dialogues, and actions), we focus on analyzing interactions (e.g., dialogue) among the characters to build such social networks. Most importantly, these relationships among characters can be extracted into similar scenes. Thereby, in this paper, we propose a novel method clustering characters using their sentimental similarities. If a minor character has a similar emotion vector to the main characters, then we assume that the minor character can be classified as a tritagonist who is helping the main character. Conversely, this minor character can be assumed to be clustered into another group and denoted as an antagonist. To evaluate the proposed approach, we show the efficiency of our proposed method by experiment in this paper.