Hinglish Sentiment Analysis: Deep Learning Models for Nuanced Sentiment Classification in Multilingual Digital Communication

Abstract

The study explores the efficiency of a hybrid LSTM-GRU deep learning model for sentiment analysis on Hinglish data, a hybrid language blending Hindi and English. Integrating Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) architectures, the model adeptly addresses linguistic intricacies, crucial for precise sentiment classification. Leveraging the combined strengths of LSTM and GRU, the hybrid model demonstrates improved memory retention and accelerated training convergence, leading to superior overall performance. Impressively, the model achieves an accuracy of 96.76%, surpassing comparable models, while precision and recall scores stand at 98.49% and 98.56%, respectively. The Hybrid LSTM-GRU model emerges as a cutting-edge and impactful tool in the realm of sentiment analysis on Hinglish data, showcasing its promise for practical deployment in diverse linguistic and cultural contexts.