Extraction of Relationship Between Characters in Narrative Summaries

摘要

Distinguishing relationships between people plays a vital role in understanding of narratives. We can experience different kinds of relationships like family, friendship, hostility, romantic love, etc. We know that Narratives are rich reflections of such relationships. Also extraction of relationship between characters is more helpful in understanding the person's actions, goals and behavior. In actual case relationships between people changes over time. So considering the dynamic nature of relationships will provide a better understanding of narratives. Existing research on analysis of people's relationships in the text has been limited Two simple schemes that use traditional machine learning algorithms. Relevant to this context, this paper proposes a system for extracting interpersonal relationships using deep learning concept convolutional neural network (CNN). The proposed approach produces a better performance over the competitive baseline model.