Desarrollan un método que identifica la depresión en las palabras escritas

Researchers at the Universidad Politécnica de Madrid have developed a machine learning-based method for the early identification of depression symptoms through textual analysis. This approach, which uses public data from social media platforms such as Reddit, demonstrates that linguistic signals associated with this disorder can be detected with high accuracy, offering an efficient alternative to more computationally complex methods. The study reveals that although affective features are decisive, the integration of social, syntactic, and thematic dimensions significantly enhances predictive capacity. The findings suggest that the use of organized lexical resources enables the capture of key cognitive and emotional patterns, establishing an optimal balance between performance and efficiency in the automatic detection of mental health indicators. This research is relevant to the field of open data because it validates the potential of public and accessible datasets to advance digital health. By demonstrating the feasibility of training models with freely available information, it promotes the development of scalable and accessible tools that could be integrated into educational or medical settings, fostering transparency and innovation based on public information to address global well-being challenges.

Source: expreso.ec
Published on 2024-02-24

Balancing Transparency and Privacy: The Double-Edged Sword of Open Data »

Related news