Sentiment Detection Using X-NLP: Harnessing Advanced NLP Techniques
M Vaneeta, Aryan Pandey, Ravi A, Mahammadinthiyaz, Ashwath · 2025
Sentiment analysis is a useful way to gain insights from text, helping organizations make informed decisions based on the emotions in the information they collect. This project aims to create a sentiment analysis system using Explainable Natural Language Processing (X-NLP) to clarify how sentiment predictions are made. The system offers users a look into how certain words and features affect sentiment classification. By using explainable AI techniques, it ensures that users not only get accurate sentiment labels but also understand why these labels were assigned, building trust and ease of use. The models are trained on the Sp1786 database and a larger data-set of 1.6 million tweets, applying both statistical methods (SVM) and deep learning techniques (LSTM). The SVM model achieved a precision of 0.6753, an AUC of 0.8432, and an F1 score of 0.6769, making it effective for real-world use.