Automatically Evaluating Opinion Prevalence in Opinion Summarization

This repository contains the ReviewNLI dataset, and code to compute the opinion prevalence metric, as introduced in the paper, "Automatically Evaluating Opinion Prevalence in Opinion Summarization" by Christopher Malon.

Running the opinion prevalence metric

Please see the metric directory.

ReviewNLI Data

See the data directory.

Simplification, for preprocessing or postprocessing

See the preprocessing directory.

Citation

If you use our metric or our dataset, please cite our paper:

@inproceedings{prevalence,
 author = {C. Malon},
 title = {Automatically Evaluating Opinion Prevalence in Opinion Summarization},
 booktitle = {The 6th Workshop on e-Commerce and NLP (KDD 2023)},
 year = {2023}
}