The CIID(Chinese Implicit Intent Dataset) dataset is obtained from manual collection and weibo crawl data. The original dataset contains a total of 5042 utterances, and the utterance is the text containing the user’s implicit intent, the average length of the data is 18.24 ,some examples are shown in the table.
We focused on seven intents:
1.Checking Weather
2.Shopping
3.Traveling
4.Booking Movie Ticket
5.Setting Reminder
6.Booking Hotel
| Utterance | Intent |
|---|---|
| 今天出去玩需要带雨伞不？ | 查询天气 |
| 快过年了，应该开始准备囤点年货了。 | 购物 |
| iPhone13好用吗，我的12今天凄惨牺牲了。 | 购物 |
| 世界那么大，我想去看看。 | 旅行 |
| 欢乐谷周围有住的地方吗? | 订酒店 |
Our dataset contains three files: train.txt, dev.txt, test.txt, and the quantities are 4242, 300, 500, respectively.
If you use the dataset, please cite this paper: A Prompt Learning Based Intent Recognition Method on a Chinese Implicit Intent Dataset CIID.