Matt Mireles
RSS Random

Humanoid: Accuracy and Fraud on mTurk (Case Study)

gethumanoid:

Through algorithmic analysis, Humanoid’s software identified 61 workers (8% of the workforce) engaged in fraud and banned them as a result. Once fraudsters were weeded out, mTurk workers generated 78% “per data field” accuracy.

Through applying various quality assurance and improvement techniques, Humanoid ultimately delivered an accuracy of 98% to Fooducate.

Permalink
2 notes
  1. mattmireles reblogged this from gethumanoid
  2. utonesm reblogged this from gethumanoid
  3. gethumanoid posted this
gipoco.com is neither affiliated with the authors of this page nor responsible for its contents. This is a safe-cache copy of the original web site.