10.1371/journal.pone.0114046.g003 Albert Yu-Min Lin Albert Yu-Min Lin Andrew Huynh Andrew Huynh Gert Lanckriet Gert Lanckriet Luke Barrington Luke Barrington Tag Clustering: Public Library of Science 2014 training mechanism consensus scale survey Genghis Khan 10 K kernel density estimation networked groups anomaly detection feature categorizations landscape peer feedback loop Genghis Khan Massively participant reasoning engine knowledge generation National Geographic expedition 2014-12-30 03:04:55 Figure https://plos.figshare.com/articles/figure/_Tag_Clustering_/1281734 <p>(a) Crowd generated tags (grey) within a 2 dimensional space (white) on a 5×5 grid; (b) Kernel density visualized through a color gradient where high density is represented as red and low density as blue; (c)Tag clusters are annotated with semitransparent circles of variable radii, determined as a dependent function of the normalized kernel density estimate of a cluster's center. Large radius circles represent a location of high agreement across the all participants. Satellite imagery provided courtesy of the GeoEye Foundation.</p>