Multi-lingual opinion mining for social media discourses: an approach using deep learning based hybrid fine-tuned smith algorithm with adam optimizer

Abstract

Multi-lingual opinion mining has become an important research area in natural language processing, particularly in the context of social media, as it provides valuable insights from user-generated content. In this paper, a deep learning approach based on a hybrid fine-tuned Smith algorithm with Adam optimizer (HFS-AO) is proposed for multi-lingual opinion mining. The approach can process three different languages, namely Marathi, Hindi, and English, which are collected from various web resources using a web scraping algorithm. The collected data is annotated using the Zero-shot instance-weighting technique and pre-processed to remove unnecessary noises and symbols, including data transformation, data cleaning and reduction, stop-word removal, tokenization, lemmatization, stemming, and Parts of Speech Tagging (POS). Naive Bayes vectorization with Laplace smoothing is used for text vectorization, followed by feature selection and classification using the hybrid fine-tuned Smith algorithm with Adam optimizer for polarity classification.The proposed method of multi-lingual opinion mining using a deep learning approach based on a hybrid fine-tuned Smith algorithm with Adam optimizer is effective in extracting valuable insights from social media data in different languages. The approach has demonstrated high accuracy, precision, recall, and F1-Score values, making it a useful tool for improving performance and customer satisfaction by identifying patterns and trends in public opinion.Additionally, the proposed method outperforms other existing methods, including PGM, MCM, CNN, and NBi-LSTM, in terms of computational time and performance making it a valuable contribution to the field of information management.