推荐系统离线评估的数据处理：综述
原标题：Data Processing for Offline Evaluation in Recommender Systems: a Survey
AI 摘要
该综述系统梳理了推荐系统离线评估中的数据预处理实践，覆盖从数据集选择、交互表示到多模态特征提取与训练-验证-测试划分的完整数据流水线。作者分析了协同、序列、会话、图、知识感知、多模态、联邦、跨域、对比学习及基于LLM的推荐等多种范式，并提出了统一的分类框架。实证分析发现，当前实践高度集中于支持度过滤等数据集级变换，而依赖表示的特征变换较少，且辅助信息表示与数据划分协议存在显著异质性，影响实验可比性与可复现性。
正文节选
Data Processing for Offline Evaluation in Recommender Systems: a Survey Multimodal Feature Extraction Abstract Offline evaluation is the dominant experimental paradigm in recommender systems research, enabling reproducible and cost-effective comparisons on historical interaction data. Yet, while considerable attention has been devoted to recommendation models and evaluation methodologies, the data processing decisions that precede model training have received less scrutiny. These decisions deter