StackOverflowed512/indian-farmer-negotiation-data

这是一个高质量、现实化的训练数据集，用于构建帮助印度农民在农产品市场（mandis）与交易商谈判更好价格的AI系统。数据集包含5,000个例子，语言为印地语/Hinglish（自然口语风格），覆盖30种作物和18个印度州。每个例子包含输入和输出字段，输入字段包括作物名称、州、市场价格、交易商报价、销售数量、市场趋势和农民销售紧迫性等，输出字段包括推荐价格、原因解释和农民可以使用的自然语言句子。数据集设计原则强调现实性、基于真实数据和自然语言。此外，还详细介绍了作物和州的覆盖范围、分布统计、包含的边缘案例以及使用方法。

A high-quality, realistic training dataset for building AI systems that help Indian farmers negotiate better prices with traders at mandis (agricultural markets). The dataset contains 5,000 examples in Hindi/Hinglish (natural spoken style), covering 30 crops and 18 Indian states. Each example includes input and output fields: input fields consist of crop name, state, mandi price, trader offer, quantity, market trend, and farmer urgency, while output fields include recommended price, reasoning, and a natural language sentence the farmer can use. The dataset is designed with principles of realism, grounding in real data, and natural language. It also details crop and state coverage, distribution statistics, included edge cases, and usage instructions.