淘宝网手机商品评论语料集

本数据集为淘宝网手机商品评论AI生成语料集（v1.0），由百图科技工作室基于AI模型合成生成，共计200,000条语料记录。所有数据均为AI模拟合成，不包含任何真实用户数据，不涉及个人隐私信息。语料内容模拟了淘宝网平台手机商品用户评论的表达风格，文本长度控制在2-50字范围内，覆盖正面、中性、负面三种情感倾向，以及已购买、考虑购买、观望中、纯分享、不推荐购买等多种购买意图类别。 数据集涵盖华为、小米、vivo、iQOO、荣耀、realme、中兴、传音、联想、谷歌等主流手机品牌，产品品类包含旗舰机、中端机、入门机、游戏手机、拍照手机、大电池手机、折叠屏手机、轻薄手机、商务手机、5G手机、性价比手机、学生手机等12个细分品类，评论主题涉及外观、拍照、信号、性价比、系统、续航、屏幕、性能、物流等核心维度。 数据集包含三层文件：纯文本文件（适用于大语言模型预训练、词向量训练等）、元数据文件（包含30个基础元数据字段及淘宝平台特有字段）、标注数据文件（在元数据基础上扩展至50个字段，涵盖情感极性与强度标注、实体与属性标注、用户画像模拟标注、比较关系标注、购买决策信号标注、内容真实度模拟标注等This dataset is the AI-generated corpus of Taobao mobile phone product reviews (v1.0), synthesized and generated by Baitu Technology Studio based on AI models, with a total of 200,000 corpus records. All data are AI-simulated and synthesized, without any real user data or personal privacy information involved. The corpus content simulates the expression style of user reviews for mobile phone products on the Taobao platform, with the text length controlled within 2 to 50 words, covering three sentiment tendencies: positive, neutral and negative, as well as multiple purchase intention categories such as purchased, considering purchase, waiting and watching, pure sharing, and not recommending purchase. The dataset covers mainstream mobile phone brands including HUAWEI, XIAOMI, vivo, iQOO, Honor, realme, ZTE, Transsion, Lenovo, Google, etc. The product categories include 12 sub-categories: flagship smartphones, mid-range smartphones, entry-level smartphones, gaming smartphones, camera-focused smartphones, long-battery-life smartphones, foldable smartphones, ultra-thin smartphones, business smartphones, 5G smartphones, cost-effective smartphones, and student-oriented smartphones. The review topics involve core dimensions such as appearance, photography, signal, cost-effectiveness, system, battery life, screen, performance, and logistics. The dataset includes three layers of files: plain text files (suitable for large language model pre-training, word embedding training, etc.), metadata files (containing 30 basic metadata fields and Taobao platform-specific fields), and annotated data files (expanded to 50 fields based on metadata, covering sentiment polarity and intensity annotation, entity and attribute annotation, simulated user profile annotation, comparison relationship annotation, purchase decision signal annotation, simulated content authenticity annotation, etc.)

搜集汇总

背景与挑战

背景概述

该数据集是由AI模型合成的淘宝网手机商品评论语料集，包含约20万条模拟评论，覆盖正面、中性、负面等多种情感倾向和购买意图。它提供了细粒度的标注信息，如情感强度、用户画像和比较关系，适用于自然语言处理任务的研究与开发。