Embedded Files
Data
The training corpus comprises a total of 3436 examples, taken from 730 unique reviews.
The test corpus contain a total of 2313 examples, from 506 reviews. Those are divided in two pairs of subtasks:
Opinion Term Extraction and Aspect Category Detection have 947 examples, from 253 unique reviews
Aspect Term Extraction and Aspect-Sentiment Quad Prediction contain the other 253 reviews, with a total of 986 target aspects and 1176 target quadruples
After the competition you can ask for the dataset access through this form.
Page updated
Report abuse