Multilingual Sentiment Analysis using DistilBert

In today's virtual age, studying feelings and sentiments through textual content and emojis is essential. This research article explores a way to use natural language processing (NLP) techniques to interpret feelings from textual content and emoji content. We used the distilbert-base-uncased-finetuned-sst2english version, a small variant model of BERT, to categorize feelings as positive or negative. Our experiments confirmed that the version executed extraordinarily properly throughout more than one language, with especially great effects on Indonesian textual content. Our assessment confirmed that the version has an excellent accuracy of91.50% and an F1 rating of 91.44% for Indonesian, highlighting its wonderful overall performance in this language. It additionally plays properly in English, French, German, Arabic, and Italian, even though overall performance may also range through language. These effects reveal the version's effectiveness for duties, social media monitoring, client comments evaluation, and inventory marketplace fashion assessment. We finish with a conclusion about the results of those findings.