Czech Dataset for Complex Aspect-Based Sentiment Analysis Tasks

TL;DR

This paper introduces a new Czech dataset for complex aspect-based sentiment analysis, enabling advanced tasks and cross-lingual research, supported by baseline results and high-quality annotations.

Contribution

It provides a novel Czech dataset designed for complex ABSA tasks with a unified annotation format, facilitating cross-lingual analysis and benchmarking.

Findings

High inter-annotator agreement (~90%)

Robust baseline results with Transformer models

Availability of 24M reviews for unsupervised learning

Abstract

In this paper, we introduce a novel Czech dataset for aspect-based sentiment analysis (ABSA), which consists of 3.1K manually annotated reviews from the restaurant domain. The dataset is built upon the older Czech dataset, which contained only separate labels for the basic ABSA tasks such as aspect term extraction or aspect polarity detection. Unlike its predecessor, our new dataset is specifically designed for more complex tasks, e.g. target-aspect-category detection. These advanced tasks require a unified annotation format, seamlessly linking sentiment elements (labels) together. Our dataset follows the format of the well-known SemEval-2016 datasets. This design choice allows effortless application and evaluation in cross-lingual scenarios, ultimately fostering cross-language comparisons with equivalent counterpart datasets in other languages. The annotation process engaged two trained annotators, yielding an impressive inter-annotator agreement rate of approximately 90%. Additionally, we provide 24M reviews without annotations suitable for unsupervised learning. We present robust monolingual baseline results achieved with various Transformer-based models and insightful error analysis to supplement our contributions. Our code and dataset are freely available for non-commercial research purposes.