DOI: 10.14704/nq.2022.20.8.NQ44712

Multilingual Sentiment Analysis Based on Convolutional Neural Network and Bert

Aniket K. Shahade, K. H. Walse, V. M. Thakare

Abstract

The large amount of text type data conveyed among people across various mediums has increased as a result of the development in technology. Sentimental analysis is made possible by the accessibility of such various reviews or sentiments of people’s emotions. Moreover, because there is a lack of standardised labelled data in the English, Hindi, and Marathi language area in natural language processing for sentiment classification and analysis is made significantly difficult. The most of the previous study in English, Hindi, and Marathi has focused on deep learning approach that heavily emphasise context-independent embeddings. Examples of these approaches include Word2Vec that give word for fixed appearance regardless of such contexts. In this paper demonstrated a cutting-edge classification accuracy for English, Hindi and Marathi sentiment analysis as a consequence, which greatly exceeds all models and techniques.

Keywords

Sentiment Analysis, Machine Learning, Deep Learning, Emotion, Sentiment Classification

Full Text

References