Context-Aware Product Review Sentiment Analysis With SBERT Representations And Boosted Rules Classifier

Read full text Summarise with VIRAT
Authors: B Maheshwari, Pasupuleti  Narasimha, Thumma Shiva, Peddaboina Sarika, Shivanadhuni  Shivateja
Cited by: 0
References: 16

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The rapid expansion of the global e-commerce industry, projected to exceed USD 7 trillion by 2030, has intensified the importance of customer reviews in influencing purchasing decisions, with over 80% of consumers relying on online feedback.However, manually analyzing large volumes of product reviews is inefficient, inconsistent, and impractical for timely business insights.Traditional sentiment analysis techniques often fail to capture contextual semantics and struggle with imbalanced data distributions.To address these challenges, this study proposes a robust Natural Language Processing (NLP) framework for automated product sentiment analysis using annotated review datasets.The pipeline begins with comprehensive text preprocessing and Exploratory Data Analysis (EDA) to clean, normalize, and understand review patterns.Context-aware semantic representations are then generated using Sentence Bidirectional Encoder Representations from Transformers (SBERT), which effectively capture sentence-level meaning and relationships compared to conventional feature extraction methods.To mitigate class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) is employed to create balanced training samples.Unlike baseline models such as Random Forest Classifier (RFC), Light Gradient Boosting Machine (LGBM), and Extreme Gradient Boosting (XGB), the proposed system integrates Deep Neural Network (DNN)-based feature selection with a Boosted Rules Classifier (BRC), enhancing both predictive accuracy and model interpretability.The framework classifies sentiments into Negative, Neutral, and Positive categories, enabling precise understanding of customer opinions.Experimental results demonstrate improved performance, scalability, and reduced bias, providing actionable insights that support product development, targeted marketing, and customer engagement strategies.

Fields of Study & Subjects

Sentiment Analysis and Opinion Mining Topic Modeling Hate Speech and Cyberbullying Detection

Publication details

Journal / source
International journal of engineering science and advanced technology.
Volume
26
Issue
3
Pages
423–434
Publication date
1 March 2026
Language
EN