Optimal Bidirectional Long Short Term Memory based Sentiment Analysis with Sarcasm Detection and Classification on Twitter Data

Sentiment analysis (SA) and sarcasm detection on social networking platforms have gained significant attention in recent times. The existence of sarcasm in social networking data like Twitter is a major cause of wrong categorization of sentiments. It remains a challenging issue from natural language processing (NLP) since it restricts the way of identifying the original sentiment of the people. Different feature engineering approaches are available in the literature to detect the presence of sarcasm in Twitter data. This study develops an Optimal Bidirectional Long Short Term Memory related Sarcasm Detection and Classification (OBiLSTM-SDC) using Twitter Data. The intention of OBiLSTM-SDC approach is for identifying and classifying the presence of sarcasm in social networks. The OBiLSTM-SDC technique encompass term frequency-inverse document frequency (TF-IDF) vectorizer. Further, the BiLSTM model can be utilized for the classification of sarcasm and the hyperparameter tuning of the BiLSTM model takes place via chicken swarm optimization (CSO) algorithm. The experimental assessment of the OBiLSTM-SDC model is taken place utilizing benchmark dataset and the outcomes were checked on various prospects. The simulation outcomes emphasized the supremacy of the OBiLSTM-SDC model over the existing approaches.