AmazonQAC

该数据集名为AmazonQAC，它是一个大规模的数据集，来源于亚马逊的搜索日志，包含了395百万个样本。这些样本包括用户键入的前缀序列，这些序列最终形成了搜索词，同时还包含了会话ID和时间戳。该数据集还包括诸如会话ID和时间戳等元数据，并已对个人可识别信息（PII）进行了清洗。它捕捉了真实的用户输入模式，包括非线性的序列和具有语义意义的建议。其规模达到了3.95亿个样本，适用于查询自动完成（QAC）任务。

This dataset is named AmazonQAC. It is a large-scale dataset derived from Amazon's search logs, containing 395 million samples. These samples include the prefix sequences typed by users that eventually form complete search queries, as well as session IDs and timestamps. The dataset also includes supplementary metadata such as session IDs and timestamps, and has been sanitized to remove all Personally Identifiable Information (PII). It captures real-world user input patterns, including non-linear sequences and semantically meaningful query suggestions. With a scale of 395 million samples, it is suitable for Query Auto-Completion (QAC) tasks.