A Comparison of Machine Learning Models to Prioritise Emails using Emotion Analysis for Customer Service Excellence

There has been little research on machine learning for email prioritization customer service excellence. To fill this gap, we propose and assess the efficacy of various techniques classifying emails into three degrees priority: high, low, neutral, based emotions inherent in content. It is predicted that after are classified those categories, recipients will be able to respond more efficiently provide better service. We use NRC Emotion Lexicon construct a labeled dataset 517,401 messages our proposal. Following that, train test four prominent models, MNB, SVM, LogR, RF, an Ensemble LSVC, RF classifiers, dataset. Our main findings suggest may used classify their emotional However, some models outperform others. During testing phase, also discovered LogR LSVC performed best, with accuracy 72%, while MNB classifier poorest. Furthermore, classification performance differed depending whether was balanced or imbalanced. conclude employ promising avenue should explored further.