Abstract :

The Rapid Growth Of Social Media Platforms Has Led To An Unprecedented Surge In User-generated Textual Data, Making Automated Analysis Essential For Extracting Meaningful Insights. The Documentary The Social Dilemma Sparked Widespread Global Discussions, Generating Millions Of Tweets That Reflect Diverse Public Opinions On Social Media Ethics And Regulation. Manual Analysis Of Such Large-scale Data Is Inefficient, Error-prone, And Lacks Scalability. To Address These Challenges, This Study Proposes A Comprehensive Natural Language Processing (NLP) Framework For Simultaneous Sentiment And Topic Classification. The Proposed Pipeline Begins With Systematic Data Preprocessing, Including Tokenization, Stop Word Removal, And Lemmatization, Followed By Exploratory Data Analysis (EDA) To Uncover Linguistic Patterns And Distributional Characteristics. For Deep Contextual Understanding, RoBERTabased Embeddings Are Employed To Extract High-quality Semantic Features From Textual Data. To Mitigate Class Imbalance, The Synthetic Minority Over-sampling Technique (SMOTE) Is Applied, Ensuring Balanced Representation Across Target Classes. The Extracted Features Are Utilized To Train Multiple Baseline Machine Learning Models, Including Decision Tree (DT) Classifier, K-Nearest Neighbor (KNN), And Naïve Bayes (NB), Enabling Comparative Performance Evaluation. Furthermore, A Deep Neural Network (DNN) Is Implemented As A Feature Extractor, Whose Intermediate Representations Are Leveraged By An Advanced Tree Alternating Optimization (TAO) Tree Classifier To Enhance Predictive Performance. The Integrated Framework, Termed Social Transform Deep Tree (STDT) Classifies Sentiments Into Negative, Neutral, And Positive, And Identifies Thematic Categories Such As Calls For Action, Emotional Reactions, And Key Insights. Experimental Results Demonstrate Improved Accuracy, Robustness, And Scalability, Highlighting The Effectiveness Of Hybrid Deep Learning And Transformer-based Approaches For Complex Social Media Text Analytics.

Published:

10-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

2199-2209

Section:

Articles

License:

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

How to Cite