A Crowdsourcing Quality Prediction Model Based on Random Forests
Hua Jian Lan, Yun Pan · 2019
With the advent of the era of big data, crowdsourcing platform on the Internet has been rapidly developed and popularized. As a business model that gathers the wisdom of groups, solves problems flexibly and effectively, crowdsourcing has attracted more and more attention and participation from people. However, the essential characteristics of crowdsourcing for all free organizations make it an urgent problem to ensure the quality of crowdsourcing and identify cheating workers quickly. This paper studies the strategies and prediction models for controlling crowdsourcing quality. By means of regression analysis, a trust model for the contractor to effectively complete the crowdsourcing task is constructed. According to the characteristic variables that influence the contractor to effectively complete the task, a random forests model is established to classify the contractor. In addition, the excellent integrated learning ability and generalization ability of the random forests are used to predict the crowdsourcing quality. It is found that the prediction model is more accurate than the traditional one.