Prompt-Based Approach for Czech Sentiment Analysis

TL;DR

This paper presents a novel prompt-based approach for Czech sentiment analysis, demonstrating superior performance over traditional methods, especially in low-data scenarios, through zero-shot and few-shot experiments.

Contribution

It introduces the first prompt-based methods for Czech sentiment analysis and shows their effectiveness in zero-shot and few-shot settings compared to fine-tuning.

Findings

Prompting outperforms fine-tuning in limited data scenarios.

Pre-training on target domain data improves zero-shot results.

Sequence-to-sequence models effectively handle aspect-based sentiment tasks.

Abstract

This paper introduces the first prompt-based methods for aspect-based sentiment analysis and sentiment classification in Czech. We employ the sequence-to-sequence models to solve the aspect-based tasks simultaneously and demonstrate the superiority of our prompt-based approach over traditional fine-tuning. In addition, we conduct zero-shot and few-shot learning experiments for sentiment classification and show that prompting yields significantly better results with limited training examples compared to traditional fine-tuning. We also demonstrate that pre-training on data from the target domain can lead to significant improvements in a zero-shot scenario.