Detection of Inappropriate Language on Social Media Platforms Using Machine Learning Algorithms

The widespread use of social media has introduced considerable communication benefits, but it has also enabled the dissemination of inappropriate and offensive language. To counter this issue, automated systems based on machine learning algorithms have been created to identify and reduce such conten...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:2024 Second International Conference Computational and Characterization Techniques in Engineering & Sciences (IC3TES) S. 1 - 5
Hauptverfasser: Mishra, Shri Om, Ahmer, Mohd, Mittal, Nupur, Maurya, Akhilesh Kumar, Kumar Singh, Amit, Kumar, Ashawani
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 15.11.2024
Schlagworte:
Accuracy
Bad words detection
Bidirectional control
Computational modeling
content moderation
Context modeling
Deep learning
Encoding
Long short term memory
machine learning
NLP
Random forests
social media platforms
Social networking (online)
Support vector machines
ISBN:9798350364682
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The widespread use of social media has introduced considerable communication benefits, but it has also enabled the dissemination of inappropriate and offensive language. To counter this issue, automated systems based on machine learning algorithms have been created to identify and reduce such content. This paper explores the effectiveness of different machine learning techniques in detecting offensive language on social media, using real-world datasets. Traditional models such as Support Vector Machines (SVM), Naive Bayes, and Random Forests were compared to advanced deep learning methods like Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN), and Bidirectional Encoder Representations from Transformers (BERT). The results show that deep learning models, particularly BERT, outperformed traditional models in terms of accuracy and F1-score, achieving an accuracy of 93.5%, compared to 87% for Random Forest and 85.2% for SVM. While deep learning models offer superior performance, they come with increased computational demands. These findings underscore the importance of selecting models based on application-specific needs, such as accuracy versus resource efficiency.
ISBN:9798350364682
DOI:10.1109/IC3TES62412.2024.10877429