Published 2025-06-26
license
Artículos de Investigación

Analysis of suicidal ideation expressed in social networks using a deep learning model.

DOI: https://doi.org/10.22490/unad.27452115.8599
Víctor Alfonso Guzmán Brand Corporación Unificada Nacional de Educación Superior image/svg+xml
Laura Esperanza Gélvez-García Corporación Unificada Nacional de Educación Superior image/svg+xml

The objective of this article is to analyze the suicidal ideation expressed in the comments posted on a discussion group on social media, by means of a pre-trained deep learning model BERT (Bidirectional Encoder Representations from Transformers). To approach the research, the method of knowledge discovery in databases (KDD) is used, characterized by the extraction of quality information, which allows to generate conclusions based on relationships or models identified within the data. As a result, it should be noted that the analysis of social media data indicates a high probability of negative intent with suicidal ideation in most of the messages. Benchmarking highlights the superiority of the BERT model in deep learning applications, especially in text analysis, thanks to its ability to understand complex patterns in unstructured data. Additionally, it is noted that searching for information about suicide on-line not only reveals inquiry about suicidal methods, but also the search for resources for help and support to manage emotional distress. Some keywords in social media comments significantly increase the likelihood of posting suicidal ideation in the future. The study concludes that the proposal for a web application highlights the need for improved tools to address suicide challenges in digital environments. The use of artificial intelligence and machine learning facilitates early detection and timely intervention to promote mental health

keywords: Analysis, Expression, Suicide, Social media, Deep learning (UNESCO Virtual Health Library)
license

How to Cite

Guzmán Brand, V. A., & Gélvez-García, L. E. (2025). Analysis of suicidal ideation expressed in social networks using a deep learning model. EducaT: Educación Virtual, Innovación Y Tecnologías, 5(2), 15-32. https://doi.org/10.22490/unad.27452115.8599
Almétricas
Metrics
Metrics Loading ...

PRIVACY STATEMENT: In accordance with the Personal Data Protection Law (Law 1581 of 2012), the names and email addresses managed by EducaT: Educación virtual, Innovación y Tecnologías will be used exclusively for the purposes stated by this journal and will not be made available for any other purpose or to any other individual. Manuscripts submitted to the publication are only accessible to the editorial team and external peer reviewers.

Design and implemented by