Sentiment Analysis using NLP Libraries and Machine Learning
Meherun Nessa Maria, Taoufik El Kabir, Sakiba Akter, Riasat Khan · 2024
Analysis of emotion is a technique that aids in determining a person’s emotion from their writings. The sector of opinion mining and emotion sentiment analysis has advanced significantly in recent years. Various studies have been done in this sector employing artificial intelligence techniques. Nevertheless, there still needs to be more procedures and frameworks available that can anticipate the highest level of accuracy in this area based on Bangla’s text. This paper illustrates a comparative analysis of different methods available for sentiment analysis on the Bangla dataset by implementing various Natural Language Processing tools and techniques like Natural Language Toolkit (NLTK) and Term Frequency-Inverse Document Frequency (TF-IDF) on three machine learning models (Multinomial Naïve Bayes, Support Vector Machine and Logistic Regression), TextBlob, and VADER frameworks. Despite the lack of a proper Bangla library, the most accurate methods have been found using a combination of NLTK and TF-IDF methodologies with SVM model to achieve acceptable performance.