Mobile App Reviews Emotion Dataset

本研究提出的数据集名为Mobile App Reviews Emotion Dataset，包含1112条来自移动应用评论的句子，这些句子已经被人工标注了情绪。该数据集旨在帮助研究移动应用评论中的细粒度情绪分析，为情感分类提供结构化指南，并为开发自动化情绪提取方法提供设计建议。数据集的创建采用了迭代人工标注过程，定义了明确的标注指南，并记录了情绪分类中的关键挑战。此外，还评估了使用大型语言模型自动标注情绪标注的可行性，分析了其成本效益和与人工标注数据的匹配程度。

The dataset proposed in this study is named Mobile App Reviews Emotion Dataset. It consists of 1,112 sentences sourced from mobile app reviews, which have been manually annotated with emotional labels. This dataset is designed to support research on fine-grained sentiment analysis of mobile app reviews, provide a structured guideline for sentiment classification tasks, and offer actionable design references for developing automated emotion extraction approaches. The dataset was developed via an iterative manual annotation workflow, with explicit annotation guidelines formulated and key challenges arising during emotion classification documented. Furthermore, the feasibility of employing large language models (LLMs) for automatic emotion annotation was assessed, and their cost-effectiveness as well as the degree of alignment with manually annotated data were analyzed.