国内外多电商平台口红电商产品评论语料集AI生成版

共计100万条数据集，为AI合成的多平台口红电商评论双语言语料集，包含国内平台（中文）50万条与国外平台（英文）50万条，。所有内容基于大语言模型通过可控提示词工程批量生成，内容长度严格控制在2-50字/单词范围内。数据集附带了情感分类、情感强度分级、用户意图、实体属性、用户画像模拟、比较关系识别、购买决策信号、内容真实度评分等多维度细粒度标注。所有数据均为虚拟合成，不涉及任何真实用户隐私，可安全用于模型训练与学术研究。This is a 1,000,000-entry bilingual corpus of AI-synthesized lipstick e-commerce reviews across multiple platforms. It consists of 500,000 Chinese entries from domestic e-commerce platforms and 500,000 English entries from international platforms. All content was generated in bulk via controlled prompt engineering based on large language models, with each entry's length strictly limited to 2 to 50 characters or words (for Chinese and English entries respectively). The dataset is equipped with multi-dimensional fine-grained annotations including sentiment classification, sentiment intensity grading, user intent, entity attributes, user profile simulation, comparative relation recognition, purchase decision signals, and content authenticity score. All data is virtually synthesized, does not involve any real user privacy, and can be safely used for model training and academic research.

搜集汇总

背景与挑战

背景概述

该数据集是一个由AI大模型生成的虚拟合成语料集，包含总计100万条中英文双语口红产品电商评论。其核心价值在于提供了情感、意图、用户画像等多维度细粒度标注，且不涉及真实用户隐私，适用于模型训练与学术研究。