Four new Reddit-derived datasets for mental health detection tasks are presented with inter-annotator agreement above 0.8 and reported model F1 scores of 93-99%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Benchmark Suite of Reddit-Derived Datasets for Mental Health Detection
Four new Reddit-derived datasets for mental health detection tasks are presented with inter-annotator agreement above 0.8 and reported model F1 scores of 93-99%.