A Crowd-Annotated Spanish Corpus for Humor Analysis

TL;DR

This paper introduces a large, crowd-annotated Spanish tweet corpus with humor labels and scores, aiming to facilitate computational humor research and address subjectivity in humor detection.

Contribution

It provides a new, publicly available Spanish humor dataset with multiple annotations, enabling improved humor recognition models and analysis of humor perception.

Findings

Krippendorff's alpha for annotations is 0.5710

The dataset includes 27,000 tweets with humor scores

Equal division between humorous and non-humorous tweets

Abstract

Computational Humor involves several tasks, such as humor recognition, humor generation, and humor scoring, for which it is useful to have human-curated data. In this work we present a corpus of 27,000 tweets written in Spanish and crowd-annotated by their humor value and funniness score, with about four annotations per tweet, tagged by 1,300 people over the Internet. It is equally divided between tweets coming from humorous and non-humorous accounts. The inter-annotator agreement Krippendorff's alpha value is 0.5710. The dataset is available for general use and can serve as a basis for humor detection and as a first step to tackle subjectivity.