A Study on Sentiment Analysis for Low-Resource Language with Emphasis on Khasi Language

Abstract

Sentiment Analysis is an NLP task of finding the opinion and classifies the opinion
expressed in a text according to its polarity (e.g., positive, negative or neutral).
Low resource sentiment analysis refers to the task of performing sentiment
analysis on text data with limited annotated data available. This is a common
scenario in many real-world applications, where annotating large amounts of
text data can be time-consuming, expensive, or even impossible. To overcome
this challenge, various methods have been proposed to perform sentiment analysis
with limited annotated data, such as transfer learning, multi-task learning,
unsupervised learning, and active learning. In this paper, we look into works
done on low-resource language sentiment analysis, compare the approaches no
these papers and compiling the success, challenges and pending issues on them.
This paper gives an outline of how sentiment analysis is performed and presents
a set of prerequisite before applying sentiment analysis on Khasi Text.