Fairy old are spending over 24 hours

Fairy tales are a part of children’s literature and usually consist of fictional stories generally aimed toward children-based reader.

These literatures will help children develop empathy and morale values 1. Children’s literature, particularly fairy tales will also help children to develop critical and imaginative thinking 2. Even though reading children’s literature brings many benefit, a collaborative study by DQ Institute and Nanyang Technological University that was conducted between August to December 2016 in Singapore showed children aged nine years old are spending over 24 hours a week watching movie. And the time spent will increase proportionally with the increase in age 3. In addition to disinterest to read, text-based children’s literature is sometimes too heavy to digest, especially for young readers which have not learnt to read completely. This might lead to misconception about the stories or morale values intended by the authors.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!

order now

To improve understanding and interest, animations can be introduced. But animation productions are usually very time- and labour-consuming which involve several experienced and skilful animators. Automated animation generation could be introduced but characters must be identified as characters are fundamental entities that build a literature. Thus, this research will focus on identifying character in literature. In the past, the entire body of automatic character identification research relies heavily on natural language processing (NLP) tools such as the Stanford Core NLP. Using NLP tools, useful information could be extracted. Groza and Corde 2 extracted information from the story in the form of triplets with the following structure: which is based on ontological reasoning.

Vargas et al. used feedback loops 3 to improve character identification further. In another paper, Vargas et al.

4 presented an approach in a fully automated system called Voz, which is integrated with Propp’s Narrative theory and Wordnet to identify roles. All of these approaches yield an accuracy of 40-70% and use Name Entity Recognition (NER) to identify character. However, NER is not sufficient and effective in identifying all possible characters. NER is unable to detect non-named characters in the story (e.g. a villager, a witch, etc.).

 The aim of this research is to propose enhancement on top of NER methods to correctly identify characters in a literature. Using the data-driven and rule-driven approach, combined with several NLP (Natural Language Processing) techniques such as POS (Part of Speech) tagger, parsing, and machine learning, characters in the story will be detected while keeping the accuracy and performance as high as possible. Insights from the results of the research will help animators in the future and advancing the field of machine learning.


I'm Ruth!

Would you like to get a custom essay? How about receiving a customized one?

Check it out