Research Article
A Posted Pricing Mechanism Based on Random Forests in Crowdsourcing Market
@INPROCEEDINGS{10.1007/978-3-030-21373-2_52, author={Lifei Hao and Bing Jia and Chuxuan Zhang}, title={A Posted Pricing Mechanism Based on Random Forests in Crowdsourcing Market}, proceedings={Security and Privacy in New Computing Environments. Second EAI International Conference, SPNCE 2019, Tianjin, China, April 13--14, 2019, Proceedings}, proceedings_a={SPNCE}, year={2019}, month={6}, keywords={Crowdsourcing Pricing mechanism Random Forests Machine learning Web spider}, doi={10.1007/978-3-030-21373-2_52} }- Lifei Hao
Bing Jia
Chuxuan Zhang
Year: 2019
A Posted Pricing Mechanism Based on Random Forests in Crowdsourcing Market
SPNCE
Springer
DOI: 10.1007/978-3-030-21373-2_52
Abstract
With the rapid development of the Internet, the combination of outsourcing and Internet has produced an overturning mode for labor cooperation – crowdsourcing. Crowdsourcing outsource the work that used to be done by internal staffs of a company or organization to non-specific people in a free and voluntary way, which concentrates the wisdom of public to solve difficult problems, greatly optimizes the rational allocation of human resources and thus improves the social productivity. In the environment of crowdsourcing market, how to set an “appropriate” price to recruit workers to complete a given task at a reasonable quality and cost is a key problem which restricts the development of it. Therefore, this paper proposes a posted pricing method based on the Random Forests (RF) algorithm in crowdsourcing market. The proposed mechanism is described theoretically and the actual crowdsourcing date is acquired from Taskcn by python spider firstly. Then, based on these empirical data, serval typical machine learning methods have been compared, which proves that RF is a very suitable method for posted pricing in crowdsourcing market. Finally, extensive experiments have been conducted and analysed for optimizing the parameters in RF and a set of parameters suitable for posted pricing in crowdsourcing is given to construct the corresponding RF model.