Paper Key : IRJ************654
Author: Gulshan Shevare
Date Published: 19 Oct 2023
Abstract
Emotion detection in textual data is an essential aspect that significantly contributes to our understanding of human expressions, sentiments, and reactions, particularly in the realm of communication and social media. This research aims to provide a comprehensive comparison of the efficacy of diverse machine learning and deep learning models in emotion detection using a meticulously curated dataset comprising human expressions. The scrutinized models encompass popular approaches such as Support Vector Machine (SVM), Random Forest, Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), and Long Short-Term Memory (LSTM). The study commences by preprocessing the dataset, employing crucial techniques such as data cleaning, tokenization, and TF-IDF (Term Frequency-Inverse Document Frequency) vectorization. These preprocessing steps are pivotal in preparing the data for subsequent training and evaluation. Following the preprocessing stage, each model is trained and rigorously evaluated using the preprocessed data. The evaluation is based on critical metrics including accuracy, precision, recall, and F1-score, enabling a thorough assessment of each model's performance. The attained results showcase that SVM and Random Forest models achieve commendable accuracy, showcasing their proficiency in emotion detection within the textual data. Conversely, the convolutional and recurrent neural network models display promising potential, albeit with a slightly lower accuracy compared to SVM and Random Forest. This disparity in accuracy underscores the nuanced strengths and weaknesses of each model in the context of emotion detection.
DOI LINK : 10.56726/IRJMETS45333 https://www.doi.org/10.56726/IRJMETS45333