Dr. Helena Garay, researcher and professor at the Faculty of Social Sciences and Humanities of the Universidad Europea del Atlántico (UNEATLANTICO, European University of the Atlantic), is collaborating in a study that designs a prediction model of depression intensity from short comments on social networks such as X.
Depressive disorders are a worldwide concern. According to the World Health Organization (WHO), about 264 million people suffer from depression, which is considered the second leading cause of suicide. This disorder is shown in different ways, with symptoms ranging from a lack of overall interest, to physical and mental problems. In this regard, it is crucial to have an automatic method for detecting and assessing the severity of depression.
The use of social media data, such as X messages, is a useful tool in the early detection of mental illness. Studies have shown that early warning signs of mental illness can be detected stemming from online activities. People suffering from depression do not always directly report their condition. However, they express their depressive feelings and thoughts on social media platforms, rather than sharing it with their family members and physicians.
So far, most research has focused on the binary detection of depression rather than on the intensity of depression. The aim of this study was to, therefore, design a model that can predict the intensity of depression in X’s posts. To achieve this, a dataset was created using hashtags associated with depression. These posts were manually classified into three levels of depressive intensity: mild, moderate, and severe. From this data, the FastText algorithm was applied to classify the intensity of depression, significantly improving the results obtained compared to other base models.
Further improvement was also achieved thanks to a weighted soft voting ensemble, which combines multiple models, including FastText, to further optimize the prediction accuracy and score. This makes the study a promising tool in detecting the severity of depression in social network users, providing them with an early warning and create a healthier society.
If you want to know more about this fascinating study, click here.
For further research, see the UNEATLANTICO repository.