karthik-44/semeval2023-afrisenti

Sentiment Analysis on low-resource African Languages. # semeval2023-afrisenti A low-resource sentiment analysis project for African Languages ## Introduction Due to the widespread use of the Internet and social media platforms, most languages are becoming digitally available. This allows for various artificial intelligence (AI) applications that enable tasks such as sentiment analysis, machine translation and hateful content detection. Despite its success, NLP technology is only widely available for high-resource languages such as English, while it remains inaccessible to many languages due to the unavailability of data resources and benchmarks. Further research is required to determine both the suitability of current natural language processing (NLP) techniques and the development of novel techniques to maximize the applications of such datasets. According to UNESCO (2003), 30% of all living languages, around 2,058, are African languages. However, most of these languages do not have curated datasets for developing such AI applications. We aim to work on sentiment analysis in a low resource setting for the AfriSenti-SemEval Task 2023 (afrisenti-semeval.github.io). This is an African-language sentiment analysis task. The dataset consists of Twitter tweets in 14 African languages. There are 3 subtasks in this SemEval task. We want to do the first subtask, Task A. Task A will tell whether a tweet expresses positive or negative sentiment, and if a tweet expresses two sentiments at the same time, then it will tell which sentiment is stronger. ## Problem Statement Given training data in a target language, perform Sentiment ### Analysis in a low resource setting and improve the prediction capabilities from existing baselines. Challenges associated to low resource language projects include: - Low availability of dataset - Twitter API does not support these languages - Crawling is difficult - Code-mixing (English language mixed in between African languages) - Skewed class distribution - Limited understanding of t …