Dataset cnn_stories from https://cs.nyu.edu/~kcho/DMQA/
Execute run_rouge.py to calcul the rouge score between machine made summaries and human made summaries.
./run_rouge.py data/my_generated_summaries.csv
where file my_generated_summaries.csv should contain (at least) two columns:
| Column | Content |
|---|---|
| Y | Human made summaries. In case of cnn dataset, all highlights concatenated to one bigger summary. |
| summary | Machine made summaries to be evaluated. |
The script will print the average precision, recall and f_score.