Feedback Seminar Analysis - An Introductory Approach from an Intelligent Perspective

摘要

This paper presents a mixture of Artificial Intelligence Topics: Automated Learning with Deep Learning and Natural Language Processing, as an advantage provided by intelligent and automated approaches in opinion mining. The main objective of this paper is to determine if a writing review can contain a positive or a negative opinion and to discover which are the best intelligent approach to detect this issue's objective. Data set used in our experiments contains 180 collected opinions/reviews from students related to a particular seminar. Methods used in our experiments are Multinomial Naive Bayes, Support Vector Machine with the RBF kernel, linear Support Vector Classification and Recurrent Neural Network with Long-Short Term Memory Unit. Computed metrics used to evaluate methods performance are accuracy, precision, recall and F1-score.