Automatic Sentiment Detection on Social Media Using Deep Learning

This article discusses the detection of emotions from textual data using deep learning techniques. The study compares the performance of the Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BiLSTM) models with Support Vector Machines (SVM) and Logistic Regression (LR) algorithms. The dataset consists of 16,000 training instances and 4,000 testing instances across six emotions: anger, fear, joy, love, sadness, and surprise. The results show that both BiLSTM and LSTM models outperform SVM and LR in emotion detection. BiLSTM achieves the highest accuracy at 93.167%, while LSTM exhibits superior performance in terms of average precision (93.22%), average recall (93.86%), and F1 score (93.04%). These findings demonstrate the effectiveness of deep learning techniques, particularly BiLSTM and LSTM, in accurately detecting emotions from text.