Hybrid Deep Learning Model for Accurate Detection of Irony and Sarcasm from Twitter Dataset

People use social media to know about trending information. Platforms like Facebook, Twitter etc. contain more data. Twitter contains tweets of opinions on any topic which can be viewed and retweeted by users. Sentimental Analysis (SA) deals with finding emotions behind the sentences posted on such social media. Some tweets on topics can be sarcastic where human intelligence cannot identify them. Sarcasm is a type of sentiment used to positively or negatively criticize ideas or situations. It includes implicit sentences and may not be understandable. In this paper, Hybrid Deep Learning Model for accurate detection of Irony and Sarcasm from Twitter Dataset (HDLM-ISTD) model is proposed to detect sarcasms. GloVe word embedding is employed for converting the words into vector form. Text Blob algorithm is applied for lemmatization and Natural language Tool Kit (NLTK) is used. Three layers of Convolutional Neural Network (CNN) and one layer of Long Short Term Memory (LSTM) are used for classifying tweets. The proposed scheme offers improved Accuracy, Precision, Recall and F1-Score.