Volume 18, No. 2, 2021

Enhancing Opinion Mining In Social Media Using A Multi-Edge Adaptive Deep Convolutional Neural Network Integrating Text And Emojis For User Sentiment Prediction


M.Kavitha , Dr.A.John Sanjeev Kumar

Abstract

In the social media age, understanding and evaluating user sentiment is essential for various uses, including public opinion tracking, customer feedback analysis and marketing. Since social media sites like Twitter have become well-known forums for people to share their thoughts and feelings on various subjects, precise sentiment analysis of textual material combined with emojis may provide a detailunderstanding of user sentiment. The paper introduces a novel Twitter sentiment prediction method, utilizing a Multi-Edge Adaptive Deep Convolutional Neural Network (MADCNN) and an emoji sentiment lexicon, which serves as the foundation for the sentiment analysis model. The preprocessing stage comprises stop word removal to remove noise from the text data, and we use Word2Vec for feature extraction to capture the semantic meaning of words and emojis. Our emoji sentiment lexicon strengthens the sentiment analysis by linking emojis with emotion scores. The MADCNN framework is designed to handle social media data, which includes text and emojis. It incorporates multiple edges for processing text and emoji inputs separately, allowing the model to extract nuanced sentiment information from both textual and non-textual elements of posts.We evaluate the effectiveness of the suggested method by contrasting it with existing sentiment analysis approaches, employing evaluation metrics including precision, recall, confusion matrix, f1-score, and accuracy. The results indicate that the suggested MADCNN performs more than conventional techniques. The suggested MADCNN, in conjunction with the emoji sentiment lexicon, provides a strong tool for predicting user opinions.


Pages: 3048-3061

Keywords: User Opinions, Sentiment Analysis, Social Media, Multi-Edge Adaptive Deep Convolutional Neural Network (MADCNN)

Full Text