Biodiversity informatics is at a juncture where public engagement in the science is inevitable. Crowdsourcing has emerged as a solution with the possibilities of utilizing intellectual resources of the mass for scientific purposes. In this study, the possibility of conducting transcription of primary biodiversity data using crowdsourcing has been explored. A crowdsourcing application for transcribing image labels from biological specimens was implemented and the accuracy of tran- scription from citizen scientists was assessed. The consensus among the users’ inputs was computed using Multiple Sequence Algorithms and then evaluated against the ground truth. The results demonstrated a mixed level of accuracy for different input fields in tran- scription. It indicates that the level of accuracy in transcriptions depends on the level of interpretation required from the participants. Further, it indicates that implementation of suggestive features in the application improves accuracy.
CITATION STYLE
Shah, M. (2014). Accuracy Assessment of Crowdsourced Data in Biological Specimen Transcription. Computer Science, (June). Retrieved from http://epublications.uef.fi/pub/urn_nbn_fi_uef-20140793/urn_nbn_fi_uef-20140793.pdf
Mendeley helps you to discover research relevant for your work.