A Study on Different Methods in Sentiment Analysis from Text

Sentiment analysis (SA) research has predominantly focused on monolingual SA over the past four years. However, there is a growing need to explore SA in languages with mixed code and expand analysis to underrepresented languages. This study aims to address this gap by investigating the creation of a gold standard dataset for sentiment analysis in mixed code languages and exploring techniques for underrepresented languages. The objective is to understand the construction of high-quality, standardized datasets encompassing mixed code languages, serving as benchmarks for evaluating SA models in languages with diverse code-mixing patterns. Currently, resources for SA in mixed code languages are limited, and comprehensive datasets will facilitate advancements in SA research. The study explores sentiment analysis in languages that have received limited attention, especially those involving code-mixing scenarios. Analyzing sentiment in these languages enhances the applicability of SA in diverse linguistic contexts and provides insights into sentiment expression and perception across cultures and language combinations. Through an extensive literature review, the study highlights the need for expanded SA research beyond monolingual analysis. By addressing the formation of gold standard datasets for mixed code languages and exploring SA techniques in underrepresented languages, this research pushes the boundaries of SA and extends its applicability to complex linguistic landscapes. The findings will contribute to advancing SA in diverse language contexts and provide a foundation for future research in sentiment analysis across multiple languages.