Robust clustering of data collected via crowdsourcing
Alba Pagès-Zamora, Georgios B. Giannakis, Roberto López-Valcarce, Pere Gimenez-Febrer
Abstract
Crowdsourcing approaches rely on the collection of multiple individuals to solve problems that require analysis of large data sets in a timely accurate manner. The inexperience of participants or annotators motivates well robust techniques. Focusing on clustering setups, the data provided by all annotators is suitably modeled here as a mixture of Gaussian components plus a uniformly distributed random variable to capture outliers. The proposed algorithm is based on the expectation-maximization algorithm and allows for soft assignments of data to clusters, to rate annotators according to their performance, and to estimate the number of Gaussian components in the non-Gaussian/Gaussian mixture model, in a jointly manner.
BibTeX
@inproceedings{icassp2017_robustclustering,
title = {Robust clustering of data collected via crowdsourcing},
author = {Alba Pagès-Zamora and Georgios B. Giannakis and Roberto López-Valcarce and Pere Gimenez-Febrer},
booktitle = {ICASSP 2017},
year = {2017}
}