Text Mining and Emotion Classification on Monkey pox Twitter Dataset A Deep Learning Natural Language Processing NLP Approach

Aim:

         To conduct an in depth analysis of emotions expressed by individuals on social media in response to the monkey pox outbreak.

Synopsis:

       The presented work outlines a research study focused on the emotion classification of social media responses to the monkey pox outbreak. The primary objective is to explore and understand the emotional responses of individuals on social media regarding the monkey pox outbreak and demonstrate the potential impact of emotion classification in enhancing our knowledge and response to the disease. The study aims to provide real time information, identify critical concerns and contributes to the public health interventions. The methodology involves extracting and preprocessing a large dataset. After preprocessing the text data, including URL and mention handling, the script explores the dataset through word clouds and charts. Feature engineering involves creating a bag of words and generating TF-IDF representation. The data is split into training and testing sets and a long short term memory (LSTM) model is trained for emotion classification, with the option for SMOTE oversampling. The trained model is evaluated on test and conducted performance evaluation

Proposed System:

            In the proposed method, emotion analysis and classification on monkey pox twitter dataset using Natural Language Processing has conducted. The proposed method begins with preprocessing of raw tweet dataset. The preprocess steps contains the following functions removal hashtag, URLs, Mentions, special characters, spell correction, replacing specific word and adding part of speech and so on. After preprocessing the text data, feature engineering involves creating a bag of words and generating TF-IDF representation. The data is split into training and testing sets and a Long Short Term Memory (LSTM) model is trained for emotion classification, with the option for SMOTE oversampling. The trained model is evaluated on the test set and accuracy is calculated along with a confusion matrix. This comprehensive approach analyses emotions expressed on Twitter during the Monkey pox outbreak, providing valuable insights for understanding public sentiment.