Text Mining for Fine-Grained Emotion Detection

Text mining for annotated Punjabi text data is done to extract the emotions of human beings. This is done to process the unstructured emotional data collected from various online resources in Punjabi and derive meaningful emotional insights from that data. The very initial phase in developing an emotion detection system is database creation. The next step in this research work is text pre-processing. Text pre-processing reduces data into a single form from multiple forms and makes the data cleansed, sorted, and structured. Feature extraction and feature selection from text data are another important step in this proposed system. Various classification models are implemented in the database to find the best classifiers that achieve an average accuracy of 81.56%. These results show the overall performance of the developed system is good and satisfactory.