RuSentNE-2023: Evaluating Entity-Oriented Sentiment Analysis on Russian News Texts

TL;DR

This paper presents RuSentNE-2023, an evaluation of targeted sentiment analysis on Russian news texts, introducing a new dataset and benchmarking various models including ChatGPT's zero-shot performance.

Contribution

The paper introduces a new annotated dataset for entity-oriented sentiment analysis in Russian news and provides benchmark results, including ChatGPT's zero-shot performance.

Findings

Best model achieved 66% Macro F-measure.

ChatGPT's zero-shot answers reached 60% F-measure.

Evaluation organized via CodaLab framework.

Abstract

The paper describes the RuSentNE-2023 evaluation devoted to targeted sentiment analysis in Russian news texts. The task is to predict sentiment towards a named entity in a single sentence. The dataset for RuSentNE-2023 evaluation is based on the Russian news corpus RuSentNE having rich sentiment-related annotation. The corpus is annotated with named entities and sentiments towards these entities, along with related effects and emotional states. The evaluation was organized using the CodaLab competition framework. The main evaluation measure was macro-averaged measure of positive and negative classes. The best results achieved were of 66% Macro F-measure (Positive+Negative classes). We also tested ChatGPT on the test set from our evaluation and found that the zero-shot answers provided by ChatGPT reached 60% of the F-measure, which corresponds to 4th place in the evaluation. ChatGPT also provided detailed explanations of its conclusion. This can be considered as quite high for zero-shot application.