Interpretability of machine learning techniques for text sentiment analysis

• Grant number: 21/02599-3
• Support Opportunities: Scholarships in Brazil - Master
• Effective date (Start): April 01, 2021
• Effective date (End): February 28, 2022
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: André Carlos Ponce de Leon Ferreira de Carvalho
• Grantee: Maynara Natalia Scoparo
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated research grant: 13/07375-0 - CeMEAI - Center for Mathematical Sciences Applied to Industry, AP.CEPID

Abstract

Sentiment Analysis is a fundamental task for understanding people's opinions regarding subjects, people or products, allowing to improve processes in different social and commercial segments. With the use of Machine Learning, it was possible to process a massive amount of data and achieve promising results. However, the models used, in their majority, are considered black box models, because they allow the minimum or nothing of understanding about its internal functioning. Thus, certain decisions are impossible to be interpreted, hence limiting the improvement of the models and understanding of the data. This project aims to explore the various techniques of Machine Learning to analyze feelings and apply methods of interpretability in order to make the models more robust. (AU)