Personalized emotion analysis aims to account for individual differences in emotional expression by adapting predictions to each writer.Existing personalization methods typically require dedicated feature extraction modules or modifications to model architectures, increasing model complexity. In this paper, we propose a prompt-based personalization framework for emotion analysis that incorporates writer information into the input prompt of a large language model (LLM), without requiring fine-tuning or architectural changes.Specifically, we investigate two complementary types of writer information: personality traits representing stable writer characteristics and post history representing writer-specific patterns of emotional expression. Experimental results on the WRIME corpus in Japanese show that incorporating temporally close post history consistently improves emotion intensity estimation across multiple LLMs. Personality information alone provides limited improvements, but yields additional gains for several models when combined with post history. These findings suggest that prompt-based personalization is a practical alternative to architecture-level personalization for LLM-based emotion analysis.

We manually construct and publicly release a Japanese dataset for Aspect-based Sentiment Analysis (ABSA), annotated with both sentiment polarity and the emotional intensities for Plutchik’s eight emotions. Existing datasets for Japanese ABSA only handle sentiment polarity classification. Therefore, we manually annotated Plutchik’s eight emotions with a four-point scale and sentiment polarity with a five-point scale to words in the Japanese sentiment analysis corpus WRIME. Analysis of this corpus revealed that word-level emotions more strongly reflect the reader’s objective impression than the writer’s subjective perspective. Furthermore, the results of evaluation experiments on word-level emotion estimation quantitatively demonstrated that while Large Language Models achieve high performance, they struggle with the estimation of the “trust” emotion. Additionally, we demonstrated that multi-task learning, utilizing both word and sentence levels, can improve performance on difficult-to-estimate subjective emotions.