Twitter Sentiment Analysis using Ensemble Based Deep Learning Techniques for Analyzing COVID-19 Tweets

The COVID-19 is a transportable disease which is rapidly spread and numerous peoples are endures and died from this disease. This paper proposed an ensemble based deep learning techniques for examining COVID-19 tweets. The COVID19 Tweets dataset is utilized in this research which consist of 3100 tweets. The preprocessing is accomplished by including the activities like stemming, tokenization and stop words removal. The preprocessed data features are extracted through Team Frequency-Inverse Document Frequency (TF-IDF). Then, it classified through ensemble classifiers such as Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU) and Bidirectional RNN (BRNN). The model performance was evaluated through accuracy, precision, recall, sensitivity and f1-score. The obtained result shows that the ensemble classifier attains high accuracy of 0.985, precision of 0.982, recall of 0.971, specificity of 0.974 and f1-score of 0.978 which is comparatively higher than existing techniques like Extra Tree Classifier (ETC) and Heterogeneous Support Vector Machine (H-SVM).