京东食品饮料评价中英双语平行语料集

本数据集为虚拟生成的中英双语平行语料集，模拟了京东平台的食品饮料评价内容风格。数据集包含10万条有效数据，每条数据同时提供地道的中文和英文版本，内容长度严格控制在中文2-120字、英文1-80词范围内。涵盖100款常见食品饮料产品，包含好评、中评、差评三种情感倾向，并附有细粒度自动标注，包含语料类型、用户画像、适用场景、传播力评分、情感强度、翻译质量等高价值维度。所有数据均为虚拟生成，不包含任何真实语料或个人数据，不涉及个人隐私，版权清晰，使用无忧。This dataset is a synthetically generated Chinese-English parallel corpus that mimics the review style of food and beverage products on the JD.com platform. The dataset includes 100,000 valid entries, each featuring native-like Chinese and English versions. The content length is strictly constrained: 2 to 120 Chinese characters for the Chinese segment, and 1 to 80 words for the English segment. It covers 100 common food and beverage products, with three sentiment orientations: positive, neutral and negative reviews. Each entry is equipped with fine-grained automated annotations covering high-value dimensions such as corpus type, user profile, applicable scenarios, virality score, sentiment intensity and translation quality. All data in this dataset are synthetically generated, containing no real corpus or personal information, and no personal privacy is involved. The copyright status is clear, ensuring worry-free usage.

搜集汇总

背景与挑战

背景概述

该数据集是一个由大语言模型虚拟生成的中英双语平行语料集，模拟了京东平台的食品饮料评价风格，包含10万条涵盖好评、中评、差评的数据，并提供了细粒度的自动标注。它适用于情感分析、机器翻译等多种自然语言处理任务，且不包含任何真实的用户隐私数据。