Identification of the level of agreement between annotators according to the guidelines for recognition of a possible depressive profile (PDP)

• Grant number: 21/08681-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: December 01, 2021
• End date: May 31, 2022
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Jair Borges Barbosa Neto
• Grantee: Victoria Caroline Cruz Alves
• Host Institution: Centro de Ciências Biológicas e da Saúde (CCBS). Universidade Federal de São Carlos (UFSCAR). São Carlos , SP, Brazil

Abstract

The mental health panorama around the world has received increasing attention due to considerable increase in past decades in emotional and psychiatric disorders diagnoses. Therefore, not only the scientific health literature, but also literature from other knowledge areas have proposed to study the subject in more depth, with the purpose of thinking about multidisciplinary strategies and interventions in mental health care. Considering a more specific perspective, the college students have been characterized as one of those with the most reports of emotional suffering, not uncommonly related to academic life factors and demands. According to the 2019 FONAPRACE report, 83.5% of students at the Brazilian federal educational institutions reported having experienced emotional difficulties in the last year. Multidisciplinary interventions are well known as effective methods in solving different demands. Thus, technology has also contributed greatly to health care in a complementary way. This study scopes technology application in the college students emotional distress identification, focusing on anonymous posting from college students, using as support the term adopted here as "possible depressive profiles" (PPD) by Online Social Network (OSN).This research is part of an interdepartmental project that will work on a computational solution creation, specialized in intervening with people who fit the possible depressive profile. This study's aim is to analyze the agreement level between evaluators regarding the symptoms in the posts in order to ensure a satisfactory degree of reliability for the machine learning (MA) model.(AU)