Overview of eRisk 2022: Early Risk Prediction on the Internet (Extended Overview)

This paper provides an overview of eRisk 2022, the sixth edition of this lab, at the CLEF conference. Since its inception, the primary purpose of our lab has been to investigate topics concerning evaluation techniques, effectiveness metrics, and other processes connected to early risk detection on the internet. Early warning models can be employed in a range of contexts, including health and safety. This year, eRisk proposed three tasks. The first one was to discover early indicators of pathological gambling. The second task was to identify early signs of depression. The third required participant to automatically complete an eating disorders questionnaire (based on user writings on social media).

keywords: early risk detection, pathological gambling, depression, eating disorders