Classification and Pattern Extraction of Stories Rated as " Tearful" Based on Narrative Structure Analysis

Abstract

In recent years, automatic generation based on narrative structure analysis has been realized in various narrative genres. However, there have been few attempts at structural analysis and automatic generation of "tearful" narratives that move readers to tears. In this study, the narrative features necessary for the automatic generation of "tear-jerking" stories was extracted. First, the stories that were evaluated as "tearful" by many people based on the voting sites on the Web were selected. Next, structural analysis for selected narrative works and categorization of tear-inducing techniques for "tear-inducing" scenes were conducted. In addition, extraction the characteristics of "tearful" stories by comparing the selected stories with general works of the same genre was attempted. As a result, it was found that the "tearful" scene tended to be associated with the presentation of hidden information by the characters and the expression of determination to overcome difficulties. It is thought that a structure that induces empathy in the user is important for a " crying" story. It is expected that the results of this research can be applied to the automatic generation of narratives to create "tear-jerking" stories.