Tasks
All tasks consist of the detection of mental disorders in users based on their comments posted on Telegram. Given a history of messages about a user, the goal is to identify whether the user suffers from the disorder or not and the context that influences the mental health problem.
Task 1. Disorder detection
Multiclass classification
Detect if the user suffers from depression or anxiety, or if there is no detected disorder at all. This is a multiclass task with three possible labels: depression, anxiety or none. Let’s see some examples:
depression: according to the WHO, depression is characterised by persistent sadness, low mood, and a lack of interest or pleasure in activities that were previously rewarding and pleasurable. A user is considered to be suffering from depression when he/she expresses everyday situations, desires, or actions related to the suffering of such pathology.
User1 (messages):
“Hola , estoy realmente mal , no se que hacer con mi vida“
“Porque las cosas no pueden acabar bien”
anxiety: the anxiety disorders are recognized by feeling intense, excessive and persistent worries, restlessness and fears about daily situations. A user is considered to be suffering from the disorder when he/she expresses everyday situations, desires or actions related to the suffering of such pathology.
User2 (messages):
"Yo tengo fobia a las agujas y tooodos los días me ponen la vacunación en los telediarios , lo estoy pasando genial tb"
"Es que somos sufridores inútiles , eso es de 1 ° de ansiedad"
none: The user does not present evidence of suffering from none of the before disorders.
Note: There may be users who present in their messages symptoms of anxiety and depression, however there is one more predominant than the other which is the one to be indicated.
Task 2. Context detection
Two-level multiclass
Same as Task 1 but adding, in the case a mental problem is detected, which is the context where the problem seems to come from. Available contexts are: Addiction context as "addiction", Emergency context as "emergency", Family context as "family", Work context as "work", Social context as "social" and Other context as "other". If no context is detected, "none" is sent. Contexts are only necessary in case the subject is predicted with depression or anxiety.
Let’s see some examples:
Addiction context: disorder is influenced by the presence of an addiction disorder, such as substance use, pathological gambling, alcoholism, among others.
User1 (messages):
“Gracias por el dato muy interesante“
“Me da ansiedad por la abstinencia de querer drogarme”
Emergency context: disorder is influenced by exceptional external factors such as pandemics, war conflicts, natural disasters, among others.
User2 (messages):
"Hola a todas y todos Me llamo José y vivo con ansiedad "
"Hum , pues últimamente peor , con esta caca con ojos del covid he sufrido un empeoramiento "
Other context: If you detect a context that is not among the previous ones.
Note: There may be users who present in their messages symptoms of anxiety and depression, however there is one more predominant than the other which is the one to be indicated.
Note: Only contexts sent with the firts positive prediction of a subject are evaluated due to the aim of the task which is early detection.
Task 3. Suicidal ideation detection
Binary classification
Detect if the user is manifesting symptoms of potential suicidal ideation. Labels will be 0 for “control” (negative, the user does not suffer from potential suicidal ideation) or 1 for “suffer” (positive). Let’s see some examples:
suffer: Suicidal ideation is characterized by a thought about not wanting to want to live having or not planning to take one's own life. A user is considered to be suffering from the pathology when he/she has knowledge on the subject and applies it in his/her daily life: he/she expresses daily situations, wishes or actions related to suffering from the pathology.
User1 (messages):
"Yo tengo fobia a las agujas y tooodos los días me ponen la vacunación en los telediarios , lo estoy pasando genial tb"
"yo tengo anorexia y no es nada lindo, me gustaría poder comer sin sentir culpa como lo hacía antes"
control: The user does not present evidence of suffering these symptoms.
Note: Only test data is provided for this task, so participants will have to train their systems with their own or external data.