| Popis: |
Sentiment Analysis examines the predominance of sarcastic language and its difficulties detecting sentiment in the text. The identification of sarcasm in the text is the focus of automatic sarcasm detection. Sarcasm recognition has been more popular in recent years and has extensive use in sentiment analysis. In this paper, we proposed sarcasm detection using machine learning. In the first phase, data has collected from social media sources such as Twitter and other synthetic datasets. The policy-based data filtration technique is used for data pre-processing and generating the normalized data vectors. The different feature extraction and selection approaches have been used for hybrid feature selection, such as TF-IDF, NLP features, Dependency features and lexicon- based features from the entire context. The various machine learning classification algorithms have been used to predict positive, negative and neutral sarcasm. The WEKA 3.7 machine learning framework has been utilized for classification. As a result, SVM produces higher classification accuracy of 95.60% over the conventional machine learning classifier. |
