ProMapCz, ProMapEn

ProMapCz和ProMapEn是由查尔斯大学数学与物理学院创建的两个产品映射数据集，分别包含1495和1555对来自不同电商网站的捷克语和英语产品。这些数据集通过手动筛选和数据预处理，包含了产品的图像、文本描述及详细规格，是目前最全面的产品映射数据集之一。数据集中的非匹配产品经过两阶段筛选，分为接近非匹配和中等非匹配两种类型，使得数据集更具挑战性，适用于训练能够区分相似但非匹配产品的模型。这些数据集主要应用于电商领域，解决产品映射和价格比较等问题。ProMapCz and ProMapEn are two product mapping datasets created by the Faculty of Mathematics and Physics, Charles University. They respectively contain 1495 and 1555 pairs of Czech and English products sourced from different e-commerce websites. These datasets have undergone manual screening and data preprocessing, and include product images, text descriptions and detailed specifications, making them one of the most comprehensive product mapping datasets currently available. The non-matching products in the datasets have been screened through a two-stage process and categorized into two types: near non-matching and medium non-matching. This design makes the datasets more challenging, making them suitable for training models that can distinguish similar but non-matching products. These datasets are mainly applied in the e-commerce field to solve problems such as product mapping and price comparison.