Article
NEUROLINGO: MULTILINGUAL SENTIMENT ANALYSIS USING RECURRENT NEURAL NETWORKS
Sentiment analysis plays a crucial role in understanding public opinion on social media platforms like Twitter, where users express views in multiple languages that reflect global market dynamics. Realtime sentiment classification across diverse languages enables businesses and policymakers to extract actionable insights, highlighting the need for robust systems capable of handling linguistic diversity and evolving temporal patterns. A key challenge lies in accurately identifying sentiment from multilingual tweets in real time, accounting for variations in language (e.g., English, Spanish, French) and cultural differences in sentiment expression, while maintaining scalability and precision. Traditional sentiment analysis methods often rely on language-specific, rule-based, or lexicon-driven approaches using predefined sentiment dictionaries, usually paired with simple classifiers like Naive Bayes. These methods are inefficient for multilingual datasets, requiring separate models per language, and they struggle with context, sequence, imbalanced data, and cross-linguistic sentiment interpretation, resulting in inconsistent performance in global applications. The proposed system, is a Python-based application with a Tkinter GUI that processes multilingual tweets. It uses NLTK for preprocessing (including case normalization, stopword removal, and stemming), TF-IDF for feature extraction, and classification models such as Decision Tree, Random Forest, and Multilayer Perceptron (MLP). Despite the title, RNN has not been implemented. The system works with a small dataset of 17 tweets in English, Spanish, French, German, and Italian, each labeled with a sentiment score from 1 to 5 stars. The MLP classifier achieved an accuracy of 87.89%, demonstrating the potential of the system for multilingual sentiment monitoring. While this tool marks progress in multilingual sentiment analysis and offers a user-friendly interface for real-time global insights, its effectiveness is limited by the small dataset and the absence of RNN architecture. Future improvements involving larger datasets and RNN integration could significantly enhance its performance and real-time capabilities for global market applications
Full Text Attachment
