Recognizing Sentiments of Maharashtra State Regional Language Marathi Text from Social Media Using Machine Learning techniques

Abstract

In the era of digital transformation, where every multinational business organisation, governing bodies and governments are trying to connect the world via a common thread. They are also trying to make their presence glocal i.e. global and local at the same time. For this one has to adapt to  regional language and its untouched population, this untouched population is majorly interested in all the relevant and updated contents in their first or second language. With this understanding their sentiments through regional language feedback, reviews and their social media opinion will play a key role in widening the business scope in the regional language geographical demographics. This study is trying to recognize and understand the regional language sentiments through a traditional approach wherein we are considering language specific stopwords, bigrams, trigram and phrases. These language specific semantic and syntactic constraints and contents contribute significantly towards recognising regional language sentiments since every language has different ways of conveying sentiments like happiness, sadness or sarcasm for that matter. The regional language is significantly modelling the way business processes operate and the way standard processes are defined due to its regional stakeholders breaking the language barrier. Regional language sentiment analysis can transform the business processes beyond just digital transformation, rather it will transform them globally and omnidirectionally through regional transformation processes. This study is focusing on how regional language sentiment analysis will contribute in today's era where Large Language Models and Generative AI are dominating the internet based products and services. This work is also trying to implement the concepts like role of bigrams, trigrams and phrases in the regional language sentiment analysis, and aims to implement new terms in Marathi language sentiment analysis like Named Entity Recognition.