tppllm/amazon-review

亚马逊评论数据集包含了从2018年1月1日至2018年6月30日的亚马逊评论。该数据集包含2245个序列，共有127054个事件，跨越18个类别类型。原始数据可在[Amazon Review Data](https://nijianmo.github.io/amazon/)获取，并在该页面上提供了引用信息。数据预处理的具体步骤可以在[TPP-LLM论文](https://arxiv.org/abs/2410.02062)和[TPP-LLM-Embedding论文](https://arxiv.org/abs/2410.14043)中找到。2025年10月28日的更新中，增加了三个时间戳字段（`timestamp_event`、`timestamp_since_start`、`timestamp_since_last_event`），以秒为单位，未做其他更改。

The Amazon Review Dataset contains Amazon reviews from January 1, 2018, to June 30, 2018. It includes 2,245 sequences with 127,054 events across 18 category types. The original data can be found at [Amazon Review Data](https://nijianmo.github.io/amazon/) with citation information provided on the page. Detailed data preprocessing steps used to create this dataset can be found in the [TPP-LLM paper](https://arxiv.org/abs/2410.02062) and [TPP-LLM-Embedding paper](https://arxiv.org/abs/2410.14043). An update on October 28, 2025, added three timestamp fields (`timestamp_event`, `timestamp_since_start`, `timestamp_since_last_event`) in seconds, with no other changes made.