DATA ANNOTATION METHOD AND SYSTEM BASED ON DATA MINING AND CROWDSOURCING
作者： Xinyu Yang , 杨新宇 , 邱楠 , 王昊奋 , Nan Qiu
DOI:
关键词:
摘要: A data annotation method based on mining and crowdsourcing, comprising: acquiring raw to be annotated (S101); performing classification crowdsourcing distribution the by using an integrated algorithm (S102); a result automatic review same algorithm, filtering marking questionable (S103); outputting that has gone through review, including (S104). Questionable results having possible problems are searched for found from among all of results, marked facilitate reviewing modifications thereof, thereby significantly facilitating finding increasing quality outputted results. The organically combines technique platforms, thus enabling massive amount accurately generated costs effectively reduced at time.
lens.org 参考文章(6)
Mark Lino Nielson, Thomas W. Kuehnel, Hiep V. Khuu, Wireless beacon filtering and untrusted data detection ,(2012)
Suraj Poozhiyil, Moe Khosravy, Max Uritsky, Vijay Mital, Robert Fries, Business application publication ,(2010)
Honglei Zhuang, Joel D. Young, Leveraging annotation bias to improve annotations ,(2014)
Sang Li, Zhang Xiaoliang, Yi Qian, Li Junwei, Liu Dongmei, Wang Jing, Xiao Hui, Jing Ming, Wang Lin, Chi Jianfeng, Wang Wenjing, Pan Yong, Wang Haipeng, Crowd-sourcing mode-based traffic information push service method and system ,(2014)
Zhou Min, Wang Xiaojun, Chou Jiangtong, Jin Cihang, Peng Tao, Method and device for rectifying marking results ,(2013)
Zhuang Yueting, Wang Donghui, Lin Yuetan, Huang Pengcheng, Xiong Kui, Hong Gaofeng, Li Ya Nan, Liang Jianzeng, Xu Can, Sun Huan, Method for integrating crowdsource annotation data based on task difficulty and annotator ability ,(2015)