Scholarlink

With the outbreak of COVID-19, we understand how social media played a crucial role by providing a platform as a means of emotional outlet for peer support and relief in the event of a health crisis. Nowadays there has been an exponential growth in the number of complex documents, user-generated data, and texts on social media, mostly on Twitter and Facebook. Social media serves as a common platform, where people share their opinions, experiences, thoughts, and precautions on COVID-19. The main purpose of this work is to find the fine-grained sentiment analysis from COVID-19 tweets from Twitter by using NLP (Natural Language Processing) techniques that classify the feature sets. User sentiment about COVID-19 tweets is determined to be positive or negative based on extracting opinions at deeper levels, that is generating a sentiment based on Polarity and Subjectivity using Textblob, rather than assigning only sentiment polarity to the tweets. Feature Extraction is performed using TF-IDF and GLOVE Word Embedding techniques on Machine Learning and Deep Learning Classifiers such as Random Forest, SVC, LOG, XGBoost, DT, NB, LSTM, and BI-LSTM. Out of all the above ML and DL approaches, LSTM on GLOVE Word Embedding achieved the highest accuracy for binary-class sentiment analysis on the COVID-19 tweets dataset.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fine-Grained Sentiment Analysis on COVID-19 Tweets Using Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • P. Appalanaidu,
  • K. Deepthi Krishna Yadav,
  • P. M. Manohar

摘要

With the outbreak of COVID-19, we understand how social media played a crucial role by providing a platform as a means of emotional outlet for peer support and relief in the event of a health crisis. Nowadays there has been an exponential growth in the number of complex documents, user-generated data, and texts on social media, mostly on Twitter and Facebook. Social media serves as a common platform, where people share their opinions, experiences, thoughts, and precautions on COVID-19. The main purpose of this work is to find the fine-grained sentiment analysis from COVID-19 tweets from Twitter by using NLP (Natural Language Processing) techniques that classify the feature sets. User sentiment about COVID-19 tweets is determined to be positive or negative based on extracting opinions at deeper levels, that is generating a sentiment based on Polarity and Subjectivity using Textblob, rather than assigning only sentiment polarity to the tweets. Feature Extraction is performed using TF-IDF and GLOVE Word Embedding techniques on Machine Learning and Deep Learning Classifiers such as Random Forest, SVC, LOG, XGBoost, DT, NB, LSTM, and BI-LSTM. Out of all the above ML and DL approaches, LSTM on GLOVE Word Embedding achieved the highest accuracy for binary-class sentiment analysis on the COVID-19 tweets dataset.