Email Sentiment Classification Using Lexicon-Based Opinion Labeling

Sentiment analysis is the most common approach to analyzing the text data to identify sentiment content. Opinions are critical to businesses and organizations because they always want to find consumers from general ideas about their products and services. The range of text data is generated from social media platforms where users share information on everything anywhere. This information is about the user’s emotions, tweets, and comments. Email is the most popular and effective way of communication overall on the Internet. Today, several applications are presented in mobiles, computers, and laptops for email messaging purposes. Some researchers have developed a lot of techniques to perform sentiment analysis. In this work, sentiment analysis for email message detection using the XGB classification method is performed. The sentiment features are extracted using count vectorization and TF-IDF transformer. Sentiments are classified using opinion labeling and three classifiers including XGBoost, gradient boosting, and random forest. The proposed work is compared with existing approaches.