淘宝网平板电脑评价中英双语平行语料集

本数据集为淘宝网平板电脑评价中英双语平行语料集，由百图科技工作室基于虚拟生成技术构建，共收录 100,000 条模拟真实电商场景的中英双语平板电脑评价语录。数据覆盖正面（59.8%）、负面（16.4%）、中性（23.7%）三种情感倾向，涉及屏幕、续航、性能、外观、音质、手写笔、系统、性价比等十余个产品维度，语料来源模拟淘宝网平台真实用户评价语境。每条数据均附有 42 维精细化标注，涵盖情感标签、用户画像、使用场景、翻译质量等多层级元数据，适用于自然语言处理、情感分析、机器翻译、电商文本挖掘等领域的研究与模型训练。This dataset is a Chinese-English parallel corpus of Taobao tablet computer reviews, constructed by Baitu Technology Studio using virtual generation technology. It contains a total of 100,000 Chinese-English parallel tablet review pairs that simulate real e-commerce scenarios. The data covers three sentiment orientations: positive (59.8%), negative (16.4%), and neutral (23.7%), involving more than ten product dimensions including screen, battery life, performance, appearance, sound quality, stylus, operating system, cost-performance ratio, and others. The corpus context simulates the real user review environment of the Taobao platform. Each entry is equipped with 42-dimensional fine-grained annotations, including multi-level metadata such as sentiment labels, user profiles, usage scenarios, translation quality, and more. This dataset is suitable for research and model training in fields such as natural language processing, sentiment analysis, machine translation, and e-commerce text mining.

搜集汇总

背景与挑战

背景概述

该数据集是由百图科技工作室通过虚拟生成技术构建的淘宝网平板电脑评价中英双语平行语料集，包含10万条模拟电商场景的评论，覆盖正面、负面和中性情感，并涉及屏幕、续航等多个产品维度。每条数据均附带42维精细化标注，适用于自然语言处理、情感分析及机器翻译等领域的研究与模型训练。