Sentiment Analysis for Code-Mixed Data Using Cellular Automata with Deep Learning Models

The proposed work presents a cellular automata-based approach to sentiment analysis in code-mixed data. Our method demonstrates promising results in effectively analyzing sentiment across multilingual tweets or sentiments. By leveraging the dynamic properties of cellular automata, our model navigates the complexities of code-mixed data, where multiple languages or dialects are intertwined within the same text. Through extensive experimentation and evaluation, we showcase the robustness and efficacy of our approach in accurately identifying sentiment-bearing components in diverse linguistic contexts. The research contributes to advancing sentiment analysis techniques in the realm of code-mixed data, offering valuable insights for understanding user sentiment in multilingual communities and enhancing communication strategies in linguistically diverse environments. Our supervised classification approach produces 89% of the F1_score for Bi-LSTM without using any kind of pre-trained word embeddings or language models.