Pretraining and Fine-Tuning Strategies for Sentiment Analysis of Latvian Tweets

TL;DR

This paper explores pre-training and fine-tuning methods to enhance sentiment analysis accuracy on Latvian tweets, achieving a significant improvement over previous approaches.

Contribution

It introduces new pre-training strategies and demonstrates their effectiveness in improving sentiment classification accuracy for Latvian tweets.

Findings

Achieved 76% sentiment analysis accuracy on Latvian tweets

Pre-training strategies significantly outperform previous methods

Experimental results validate the effectiveness of the proposed approach

Abstract

In this paper, we present various pre-training strategies that aid in im-proving the accuracy of the sentiment classification task. We, at first, pre-trainlanguage representation models using these strategies and then fine-tune them onthe downstream task. Experimental results on a time-balanced tweet evaluation setshow the improvement over the previous technique. We achieve 76% accuracy forsentiment analysis on Latvian tweets, which is a substantial improvement over pre-vious work