PerSHOP

PerSHOP是由伊朗德黑兰大学的电子与计算机工程学院创建的波斯语购物对话数据集，旨在为购物对话系统提供训练数据。该数据集包含21925条语句，分布在15个不同的领域，共有1061个对话。数据收集通过众包方式进行，历时30天，由122名众包工作者完成。数据集的创建过程包括自动和手动标注，以确保数据的质量。PerSHOP数据集主要用于训练和评估自然语言理解模型，特别是在意图分类和实体提取任务中，为波斯语购物对话系统的发展提供了基础。

PerSHOP is a Persian shopping dialogue dataset created by the School of Electrical and Computer Engineering, University of Tehran, Iran, designed to provide training data for shopping dialogue systems. This dataset contains 21,925 utterances distributed across 15 distinct domains, with a total of 1,061 dialogues. The data was collected via crowdsourcing over a 30-day period, completed by 122 crowd workers. The dataset was developed with both automatic and manual annotation to ensure data quality. The PerSHOP dataset is primarily used for training and evaluating natural language understanding models, particularly in intent classification and entity extraction tasks, providing a foundational resource for the development of Persian shopping dialogue systems.