Analysis of Literary based on Deep Emotional Network

Literary works are mainly reviewed by professional judges, whose evaluation and analysis are limited by their own experiences and preferences, which makes it difficult to find the value of the works. We design a method to analyze the quality of literary works based on deep characters emotional network. Taking the representative works of Nobel Prize in Literature and nominees for literature as the research objects, we extract the entity relationship network. The node is the entity in the works, and the edge is the relationship between entities. With the help of graph embedding algorithm, we analyze the similarities and differences of different works in emotional distribution of characters. The results show that the works of the Nobel Prize in Literature winners in 2012 and 2013 have the characteristics of high concentration of characters and even emotion fluctuation.