Multilingual Sentiment Analysis for Mobile Gaming: A Comparative Study of Machine Learning and Hybrid Deep Learning Approaches
Year: 2024
Type: conference-paper
Abstract: The rapid expansion of the mobile gaming industry has underscored the need to understand user sentiment, with social media platforms like X (Twitter) providing key insights. This study applied sentiment analysis to English and Turkish tweets, utilizing the TEMSAP-CNNLSTM model a hybrid architecture combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for precise classification of complex textual data. The model’s performance was benchmarked against traditional machine learning methods, including Logistic Regression, Support Vector Machines (SVM), and K-Nearest Neighbors (KNN). Results revealed that TEMSAP-CNNLSTM consistently achieved superior performance, with the highest Accuracy, Precision, Recall, and F1-Score across both datasets. The model attained 96% accuracy on English and Turkish training datasets and 93% and 92% on English and Turkish test datasets, respectively. These findings highlighted the model’s capability in handling sentiment data, surpassing traditional approaches while demonstrating robust generalizability across languages. The TEMSAP-CNNLSTM model offers valuable insights for mobile game developers and suggests broader applicability for other industries requiring accurate sentiment analysis in multilingual contexts.
Authors: Erol Kına, Recep Özdağ
Institution: Van Yüzüncü Yıl Üniversitesi
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4405331148