Abstract :

The Rapid Expansion Of Telecom Services Has Generated Massive Volumes Of Customer Interaction Data, Creating A Critical Need For Intelligent Systems Capable Of Extracting Meaningful Insights From Conversational Transcripts. Sentiment Detection In Telecom Data Is Essential For Evaluating Customer Satisfaction, Enhancing Service Quality, And Supporting Data-driven Decision-making. However, The Problem Arises From The Unstructured And Context-dependent Nature Of Textual Conversations, Where Sentiments Are Often Implicit And Influenced By Complex Linguistic Patterns. Earlier Systems Rely On Basic Machine Learning (ML) Techniques And Manual Feature Engineering, Which Fail To Effectively Capture Contextual Semantics, Resulting In Limited Accuracy, Poor Generalization, And Reduced Scalability When Applied To Large And Diverse Datasets. These Limitations Highlight The Need For A More Robust And Adaptive Framework That Can Handle Real-world Telecom Data Efficiently. To Address This, The Proposed System Integrates Advanced Natural Language Processing (NLP) Techniques, Including Text Preprocessing, Tokenization, Stopword Removal, And Lemmatization, Along With Transformer-based Embedding Generation Using GooglePALM. These Embeddings Provide Rich Contextual Representations Of Text, Which Are Then Utilized By Multiple ML Classifiers, Including Logistic Regression (LRC), Decision Tree Classifier (DTC), Extra Trees Classifier (ETC), Boosted Rules Classifier (BRC), And Prposed FIGS (Fast Interpretable Greedy-Tree Sums) Ensemble Classifier. Additionally, Class Imbalance Is Handled Using Resampling Techniques, And Performance Is Evaluated Using Metrics Such As Accuracy, Precision, Recall, And F1-Score. The System Is Further Deployed Through A Web-based Framework Using Django Web Framework, Enabling Real-time Sentiment Prediction From Uploaded Datasets, Thereby Improving Scalability, Accuracy, And Practical Applicability.

Published:

22-6-2026

Issue:

Vol. 26 No. 6 (2026)

Page Nos:

1269-1276

Section:

Articles

License:

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

How to Cite