Authors:
R. Mohan, B. Rohith, G. Nethaji, R. Angeline, Rahul Chauhan

Addresses:
Department of Computer Science and Engineering in Artificial Intelligence and Machine Learning, SRM Institute of Science and Technology, Ramapuram, Chennai, Tamil Nadu, India. Department of Fitch Ratings, The Fitch Group, New York, United States of America.

Abstract:

To extract insights that can be put into action from the continuously increasing volume of customer-generated content on digital platforms, sentiment analysis has evolved into a vital tool. Within the scope of this research, an artificial intelligence-based customer feedback analysis system is presented. This system combines sentiment categorization with a probability-driven confidence score to generate trustworthy, interpretable predictions of customer opinion. Data collection, text cleaning and normalization, feature extraction using Bag-of-Words and TF-IDF vectorization, and sentiment classification using a collection of deep learning and machine learning models are all included in the proposed framework. These models include Random Forest, Naive Bayes, Logistic Regression, a Feedforward Neural Network, and RoBERTa. This robustly optimized BERT pre-training method estimates the confidence of each prediction and flags ambiguous or borderline reviews for further investigation. The results of an experimental evaluation show that integrating transformer-based contextual representations with probability-based confidence assessment significantly improves the explainability and reliability of automated sentiment analysis. This enables businesses across a wide variety of industries to understand their customers' preferences better, give high-confidence feedback a higher priority, and make well-informed strategic decisions about their products, services, and marketing. 

Keywords: Sentiment Analysis; Customer Feedback Analysis; Sentiment Confidence Scoring; Opinion Mining; Machine Learning; RoBERTa and Transformer; Deep Learning; Performance Evaluation.

Received on: 29/07/2025, Revised on: 24/09/2025, Accepted on: 01/12/2025, Published on: 09/06/2026

DOI: 10.69888/FTSFDS.2026.000696

FMDB Transactions on Sustainable Finance and Data Science, 2026 Vol. 1 No. 2, Pages: 84-92

• Views : 25
• Downloads : 4

Download PDF