Classification of Fine-Grained Emotions

This paper classifies the annotated Punjabi emotional text data into six predefined basic classes of emotions: anger, disgust, fear, happiness, sadness, and surprise. Various existing classifiers are implemented with the developed system, and out of them, the best classifier is selected for classification of the fine-grained annotated emotional database of Punjabi text. In total, seven existing classifiers are applied and discussed in this research work: Support Vector Machine, Artificial Neural Network, Gaussian Naive Bayes, Decision Tree, Random Forest, Logistic Regression, and Hybrid (Random Forest + Decision Trees + SVM) and tested with various statistical measures. The classification system is trained on an annotated Punjabi corpus, 30% of the data used is for system testing, and 70% is for training the developed system.