Abstract :

The Rapid Expansion Of Digital Communication Platforms Has Resulted In Massive Volumes Of Unstructured Textual Data, Making Sentiment Analysis An Essential Task For Extracting Meaningful Insights In Domains Such As Customer Feedback, Telecom Interactions, And Social Media Analytics. However, Accurately Identifying Sentiment From Text Remains Challenging Due To Contextual Ambiguity, Linguistic Variability, Sarcasm, And Implicit Expressions. Traditional Approaches Are Predominantly Manual, Where Human Analysts Interpret Text And Assign Sentiment Labels Based On Predefined Rules Or Subjective Judgment. Although These Methods Can Capture Contextual Meaning To Some Extent, They Are Timeconsuming, Inconsistent, Non-scalable, And Prone To Human Bias, Leading To Unreliable Results For Largescale Data. These Limitations Highlight The Need For An Automated, Scalable, And Context-aware System Capable Of Delivering Accurate And Consistent Sentiment Predictions. To Address These Challenges, This Work Proposes A Hybrid Sentiment Analysis Framework That Integrates Sentence Bidirectional Encoder Representations From Transformers (SBERT) For Contextual Embedding Generation, Deep Neural Networks (DNN) For Feature Extraction, And Boosted Rules Classifier (BRC) For Interpretable Classification. The System Includes Preprocessing Steps Such As Tokenization, Stopword Removal, And Lemmatization To Ensure Clean Input Data. Synthetic Minority Over-sampling Technique (SMOTE) Is Applied To Handle Class Imbalance. Additionally, Machine Learning (ML) Models Including Random Forest Classifier (RFC), Light Gradient Boosting Machine (LGBM), And Extreme Gradient Boosting (XGBoost) Are Used For Comparative Evaluation. The Proposed System Improves Accuracy, Enhances Contextual Understanding, And Provides Interpretable Results, Making It Suitable For Real-world Sentiment Analysis Applications.

Published:

22-6-2026

Issue:

Vol. 26 No. 6 (2026)

Page Nos:

1277-1286

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite