Sentiment Analysis for Automated Email Response System

Email is the most widely used form of written communication, especially for teacher-student communication. Most routine email responses are repetitive in nature but consume a lot of academic time. An automated email response system can save much of this time. This research is focused on detection and classification of sentiment in student-teacher email conversations. This study aims to find out markers of politeness and impoliteness within email. This will ultimately help in generating more human-like response. Taking into account Ekman’s basic emotions, we have proposed a method to find out these emotions using deep learning. Limiting our research to student-teacher conversations, gives us a close domain for finding sentiment and its intensity.