Tweet sentiment analysis: leveraging weighted word embedding to evaluate criticality and security of events

Currently, one of the most important sectors globally is social media. It has been reported that about 74% of people on the planet utilise social media. This has spurred a lot of social media study. Sentiment analysis of real-time social media data for security purposes is one such helpful application. In this paper, we present a weighted word embedding approach for classification of tweets related to education. The proposed approach utilised weighted GloVe word embedding with a sequential neural network (SNN) architecture to enhance the accuracy of sentiment classification. Moreover, this paper incorporates the weighted word embeddings, assigning varying weights based on their relevance to sentiment analysis. In the empirical analysis, we evaluate the predictive performance of different word embedding schemes, including SNN word embedding, GloVe, and DIM, with deep neural network architectures. The performance of the proposed approach is measured using precision, recall, F1 score and accuracy. The experimental results showcase promising performance, effectively capturing valuable insights into public opinion during the pandemic. The combination of TF-IDF weighted GloVe word embedding and the SNN architecture proves to be a robust approach for sentiment analysis, providing accurate sentiment classification reviews on Twitter.