Active Learning and Crowdsourced Datasets

Data and the MATLAB loader accompanying Scaling Up Crowd-Sourcing to Very Large Datasets: A Case for Active Learning (PVLDB 2014).

Terms of use

Please cite the paper when using any part of this dataset:

Barzan Mozafari, Purnamrita Sarkar, Michael J. Franklin, Michael I. Jordan, and Samuel Madden. “Scaling Up Crowd-Sourcing to Very Large Datasets: A Case for Active Learning.” Proceedings of the VLDB Endowment 8(2), 125–136, 2014. DOI: 10.14778/2735471.2735474.

Files

The .desc files document each dataset. The .vis files are numeric matrices; paired .crd files contain string-valued fields. CrowdManager.m loads the supplied formats in MATLAB.

Face / vision

• face4.desc
• face4.crd
• face4.csv
• face4.details
• face4.vis

Tweet sentiment

• tweets10k.desc
• tweets10k.crd
• tweets10k.details
• tweets10k.dict
• tweets10k.vis
• tweets100k.desc
• tweets100k.details
• tweets100k.dict
• tweets100k.vis (52 MiB download; served by the University archive)

For questions about these historical materials, contact [email protected].