Data Analytics for Business and Societal Challenges
Loading...
Paper Number
1879
Paper Type
Completed
Description
Crowdsourcing contests are popular mechanisms for firms to obtain solutions to tasks from external sources. Platforms allow firms to decide the attributes of their contests, such as prize and duration. Selecting an appropriate set of attributes can increase the probability of success. We develop a framework that identifies an effective prediction model, and maximizes the predicted probability of success of a crowdsourcing contest using optimization techniques. The framework considers design features of a focal task as also those of other open tasks competing with the focal task. The contest design problem is formulated as a mixed linear knapsack problem, with several novel variants. We present solution procedures for these variants, and validate the framework using a real-world dataset. Our results show that the contest designs recommended by the proposed framework can increase success probabilities significantly relative to what seekers specify.
Recommended Citation
Zhu, Wangsheng; Mo, Jiahui; Menon, Syam; and Sarkar, Sumit, "A Framework for Optimal Crowdsourcing Contest Design" (2021). ICIS 2021 Proceedings. 11.
https://aisel.aisnet.org/icis2021/data_analytics/data_analytics/11
A Framework for Optimal Crowdsourcing Contest Design
Crowdsourcing contests are popular mechanisms for firms to obtain solutions to tasks from external sources. Platforms allow firms to decide the attributes of their contests, such as prize and duration. Selecting an appropriate set of attributes can increase the probability of success. We develop a framework that identifies an effective prediction model, and maximizes the predicted probability of success of a crowdsourcing contest using optimization techniques. The framework considers design features of a focal task as also those of other open tasks competing with the focal task. The contest design problem is formulated as a mixed linear knapsack problem, with several novel variants. We present solution procedures for these variants, and validate the framework using a real-world dataset. Our results show that the contest designs recommended by the proposed framework can increase success probabilities significantly relative to what seekers specify.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
14-Data