Elements for Automatic Identification of Fallacies in Mexican Election Campaign Political Speeches
Kenia Nieto-Benitez, Noé Alejandro Castro-Sánchez, Hector Jimenez Salazar, Gemma Bel-Enguix, Dante Mújica‐Vargas, Juan Gabriel González Serna, Nimrod González-Franco · Programming and Computer Software · 2023
Abstract Political speeches frequently use fallacies to sway voters during electoral campaigns. This study presents an approach for implementing machine learning models to automatically identify a specific type of fallacy known as an “appeal to emotion” fallacy. The objective is to establish a set of elements that enable the application of fallacy mining, as in existing literature, fallacies are typically identified manually, and there is no established structure for applying mining techniques. Our method utilizes features derived from an emotion lexicon to differentiate between valid arguments and fallacies, and we employed Support Vector Machine and Multilayer Perceptron models. Our results indicate that the Multilayer Perceptron model achieved an F1‑score of 0.60 in identifying fallacies. Based on our analysis, we recommend the use of lexical dictionaries to effectively identify “appeal to emotion” fallacies.