Graph Model for Detection of text unstructured data such as Sarcasm

  • Axel Rodríguez-García Universidad Tecnológica de Panamá
  • Armando Jipsion Universidad Tecnológica de Panamá
Keywords: Unstructured Data, NLP, sarcasm, graph model

Abstract

Sarcasm is frequently characterized as verbal incongruity to communicate scorn. It is a nuanced type of language with which people express something contrary to what is suggested. Perhaps the greatest test in building frameworks to consequently recognize unstructured information, for example, mockery, is the absence of huge, commented on informational indexes. We propose a diagram-based procedure in building conservative language models for sarcasm recognition. This strategy is likewise intended to utilize little information, it could help in different regions like disdain discourse, counterfeit news, and so forth. This charting strategy permits specialists to explore different parts of NLP without obtaining a huge dataset. These days, it still remains a challenge to unmistakably distinguish human slants and feelings by utilizing AI. Associations can use a superior philosophy to settle on proactive choices in basic circumstances. A definite investigation of our examination would hoist the current content mining applications and may help understand better the effect of mockery from the customers and partners communicated in a web-based media climate. We exhibit that straightforward classifiers worked from the model can recognize mockery very well, which they sum up 5 % better than those of the cutting edge.

Published
2021-01-01
Section
Research Articles for the Regular Issue