ci25-nishio

The supplemental materials of "Rationale-aware Label Aggregation in Crowdsourcing" in CI2025.

synthetic_data

filename format:
redundancy|number of tasks per worker|average accuracy|group ratio|file number
file content format:
task, worker, label, rationale, true label

RAAG.py

Our proposed method
input: .csv
task, worker, label, rationale, true label
output:
estimated labels and accuracy