基于大语言模型的移动应用评论细粒度情感分类实证研究
原标题：Fine-Grained Emotion Classification from Mobile App Reviews: An Empirical Study with Large Language Models
AI 摘要
该研究针对移动应用评论的细粒度情感分类问题，基于Plutchik八类基本情感标注框架和1090条人工标注数据，系统比较了编码器微调（多标签与二元集成）与解码器零样本/少样本提示方法，并评估了损失重加权、重采样和生成式数据增强等类别不平衡缓解策略。结果显示，微调编码器在基线时大幅落后于最佳解码器少样本提示，但结合生成式增强和正例加权损失后可缩小差距，且推理延迟低至三个数量级，稀有情感类别F1提升显著。作者发布了实验流程、合成语料和微调检查点。
正文节选
Fine-Grained Emotion Classification from Mobile App Reviews: An Empirical Study with Large Language Models Abstract Context: Fine-grained emotion classification of mobile app reviews enables requirements engineering activities that go beyond polarity-based opinion mining, including emotionally informed issue prioritisation and feature-oriented feedback analysis. However, automatic fine-grained emotion extraction from app reviews remains understudied. Objectives: Building on a previously publishe