Lexical Event Ordering with an Edge-Factored Model

This is the data we used in the paper:

Download the bundle here.

The bundle includes preprocessed recipes files (by the Stanford parser) and all events extracted (from the parse trees) for each recipe. We provide three files in the bundle, train_set, dev_set and test_set that correspond to the way we partitioned the data for our experiments. We also provide the URLs for ZIP files that contain the recipes we used at http://www.ffts.com. See the README here.