Highlights

Abstract

Keywords

۱٫ Introduction

۲٫ Related work

۳٫ Relevance of the personal statements for depression detection

۴٫ Adapting DPP-EXPEI to the depression detection task

۵٫ Results and discussion

۶٫ Conclusions and future work

Declaration of competing interest

Acknowledgments

References

Abstract

Depression is a common and very important health issue with serious effects in the daily life of people. Recently, several researchers have explored the analysis of user-generated data in social media to detect and diagnose signs of this mental disorder in individuals. In this regard, we tackled the depression detection task in social media considering the idea that terms located in phrases exposing personal statements (i.e., phrases characterized by the use of singular first person pronouns) have a special value for revealing signs of depression. First, we assessed the value of the personal statements for depression detection in social media. Second, we adapted an automatic approach that emphasizes the personal statements by means of a feature selection method and a term weighting scheme. Finally, we addressed the task in hand as an early detection problem, where the aim is to detect traces of depression with as much anticipation as possible. For evaluating these ideas, benchmark Reddit data for depression detection was used. The obtained results indicate that the personal statements have high relevance for revealing traces of depression. Furthermore, the results on early scenarios demonstrated that the proposed approach achieves high competitiveness compared with state-of-the-art methods, while maintaining its simplicity and interpretability.

۱٫ Introduction

Depression is one of the most common mental health illnesses that affects seriously the daily life of people. It can produce a variety of emotional and physical problems that can diminish the activities of an individual provoking negative effects on her/his surrounding personal context, work [1] or school [2], and even basic human needs such as sleeping [3] and eating [4]. Severe cases can lead to self-harm and suicide [5]. Although fortunately, depression is a treatable disorder, it is often undetected due to several reasons such as the patient’s own inability to recognise the problem or because of the social stigma associated with mental illnesses. Recently, the rise in the use of social media has opened new opportunities for detecting depression [6–۸]. In these platforms, people freely share and express their thoughts and feelings. Furthermore, often these media are used by depressed users to gather information about their illness or to discuss about their problems and symptoms.

Based on the idea that language is a powerful indicator of personality, social or emotional status, as well as mental health [9], several research works have leveraged the content generated by social media users as a rich source of knowledge to study, infer, and track users with mental illnesses such as depression. Particularly, it has been demonstrated that individuals having depression manifest changes in their language and behaviour (e.g., greater negative emotion and high selfattentional focus) [10]. In this regard, the development of methods for automatic depression detection on social media has gained a special interest in the computational linguistics research community. Such a challenging task has been generally tackled as a text classification problem, considering a wide variety of text representations and classification models, and concluding that the thematic and stylistic aspects are different between depressed and mental healthy people [11].