IndoBERT-Sentiment: Context-Conditioned Sentiment Classification for Indonesian Text

TL;DR

IndoBERT-Sentiment is a novel Indonesian sentiment classifier that incorporates topical context, significantly improving accuracy over existing models by leveraging context-conditioned predictions.

Contribution

This work introduces a context-conditioned sentiment classification model for Indonesian that outperforms general-purpose models by utilizing topical context during prediction.

Findings

Achieved an F1 macro of 0.856 and 88.1% accuracy on the test set.

Outperformed baseline models by 35.6 F1 points in head-to-head evaluation.

Effectively transferred context-conditioning from relevancy classification to sentiment analysis.

Abstract

Existing Indonesian sentiment analysis models classify text in isolation, ignoring the topical context that often determines whether a statement is positive, negative, or neutral. We introduce IndoBERT-Sentiment, a context-conditioned sentiment classifier that takes both a topical context and a text as input, producing sentiment predictions grounded in the topic being discussed. Built on IndoBERT Large (335M parameters) and trained on 31,360 context-text pairs labeled across 188 topics, the model achieves an F1 macro of 0.856 and accuracy of 88.1%. In a head-to-head evaluation against three widely used general-purpose Indonesian sentiment models on the same test set, IndoBERT-Sentiment outperforms the best baseline by 35.6 F1 points. We show that context-conditioning, previously demonstrated for relevancy classification, transfers effectively to sentiment analysis and enables the model to correctly classify texts that are systematically misclassified by context-free approaches.