Dataset of Movie Reviews Annotated for Irony This dataset has been used for an irony detection task in Spanish.
The dataset is composed of movie reviews annotated with information on the presence of irony. Each sentence in every review has three annotations, and sentences with two or more irony annotations are considered ironic.
The dataset contains a total of 883 sentences, distributed as follows:
- 63 ironic sentences
- 820 non-ironic sentences
This dataset is suitable for evaluating irony detection models in Spanish.