---
license: mit
task_categories:
- text-classification
language:
- es
tags:
- fake-news
- misinformation
- satire
- spanish
- beto
pretty_name: Unified Spanish Misinformation and Satire Corpus (USMSC)
size_categories:
- 10K<n<100K
---

# Unified Spanish Misinformation and Satire Corpus (USMSC)

## Dataset Description
This dataset is a comprehensive, deduplicated, and systematically structured corpus for domain-specific misinformation detection in Spanish social media text. It addresses the critical gap in Spanish-language resources by unifying multiple distinct datasets into a single, highly refined corpus. 

Crucially, this dataset employs a **three-class formulation (Fake, Real, Satire)**. Recent literature and empirical testing demonstrate that conflating satire with malicious fake news creates a fundamental construct-validity problem, as models risk learning source-specific humor or parody markers instead of robust misinformation cues. By isolating satire into its own distinct class, this corpus enables the training of highly calibrated models (such as BETO) capable of distinguishing creative parody from deceptive content.

### Dataset Structure
The corpus consists of exactly **61,674 unique records**, structured into three distinct target classes. This structure achieves a methodological balance where Deceptive + Satire records (49.8%) effectively balance Truthful records (50.2%).

| Class Label | Category | Count | Percentage | Description |
| :---: | :--- | :--- | :--- | :--- |
| **0** | `FAKE` | 21,746 | 35.3% | Deceptive, malicious, or misleading misinformation. |
| **1** | `REAL` | 30,943 | 50.2% | Factual, verified, and legitimate news articles. |
| **2** | `SATIRE` | 8,985 | 14.5% | Parody and ironic content. |
| **-** | **Total** | **61,674** | **100%** | After rigorous deduplication and normalization. |

## Corpus Construction & Methodological Evolution
The dataset was curated by unifying four publicly available Spanish-language sources:
* **Posadas-Durán et al.:** 572 records
* **Acosta et al.:** 598 records
* **Tretiakov et al.:** 2,000 records
* **Blanco-Fernández et al.:** 57,231 records

**Preprocessing and Deduplication Protocol:**
The initial merge yielded 60,401 records showing a binary imbalance (41.3% Fake vs. 58.7% Real). To prevent source leakage and ensure robust generalization, the raw data underwent strict normalization (lowercasing, removal of non-alphanumeric characters, and whitespace standardization). Content hashing was then applied, resulting in the removal of exactly 7,712 near-duplicates. 

This process yielded 52,689 unique factual/deceptive articles (21,746 Fake; 30,943 Real). Finally, 8,985 satirical records were appended as a distinct 3rd class, resolving the construct validity vulnerability and producing the final 61,674 records.

## Applications & Baselines
This corpus was expressly designed for the fine-tuning of specialized Transformer encoders. Under identical evaluation protocols, the Spanish-specific **BETO encoder** achieves State-of-the-Art performance on this dataset, isolating the `SATIRE` class with a perfect F1-score of 1.00 and achieving an overall Macro F1 of 0.9095.

It is highly recommended to use this dataset for:
* Multi-class classification of Spanish misinformation.
* Stylometric and semantic analysis of deceptive vs. satirical language.
* Training efficient, domain-specific NLP microservices.

## Citation
If you use this dataset in your research, please cite our paper published in *Informatics (MDPI)*:

**DOI:** [10.3390/informatics13060083](https://doi.org/10.3390/informatics13060083)

### BibTeX
```bibtex
@Article{informatics13060083,
AUTHOR = {Avilés, Gabriel Hurtado and Reyes-Ortiz, José A. and Mora-Gutiérrez, Román A. and Cuevas, Josué Padilla and Alcántara, Óscar Herrera},
TITLE = {Systematic Fine-Tuning of Transformer Models for Domain-Specific Misinformation Detection in Spanish Social Media Text},
JOURNAL = {Informatics},
VOLUME = {13},
YEAR = {2026},
NUMBER = {6},
ARTICLE-NUMBER = {83},
URL = {https://www.mdpi.com/2227-9709/13/6/83},
ISSN = {2227-9709},
ABSTRACT = {While social media platforms are primary vectors for misinformation, automated detection systems remain largely confined to English. This paper presents a transferable, three-stage framework for fine-tuning transformer models to detect domain-specific deceptive content in Spanish. The pipeline comprises: (1) corpus unification, merging fragmented datasets into a 61,674-article resource mapped into three classes (Real, Fake, Satire) to prevent stylistic confounding; (2) systematic model optimization, extensively benchmarking classical metaheuristics against eight transformer architectures (including mBERT, XLM-RoBERTa, and BETO) using strong regularization to mitigate overfitting; and (3) production deployment, encapsulating the optimized model as a containerized web application for real-time inference. Through rigorous experimentation, the Spanish-specific BETO encoder emerged as the strongest model for this task, achieving 89.18% overall accuracy. The model attains a near-perfect in-source F1-score on the satire class; however, a strict source-held-out test reveals that this performance is highly source-dependent—recall on satire from an unseen outlet drops to 0.08—indicating that single-source class construction leads the model to recognize the source rather than a generalizable category. We report this finding as a central methodological result: corpus design, and in particular the source diversity of each class, is the primary determinant of whether the framework generalizes. Adversarial robustness tests using named-entity masking and typo injection provide complementary evidence on the model’s reliance on semantic versus surface cues. The methodology is designed to be adaptable across domains: by substituting the training corpus, the same framework may in principle be retargeted to other digital threats, such as investment scams and phishing, provided that suitable labeled corpora are constructed and validated for each new domain. The complete framework, dataset, and application are released as open-source resources to support reproducible research and practical countermeasures against online misinformation.},
DOI = {10.3390/informatics13060083}
}
```