Optimizing Software Crowdsourcing Requirements Design through Machine Learning

Crowdsourcing, a distributed computing model on the internet, has opened new avenues for software development. Traditional in-house software development has transformed into more decentralized approaches, where software tasks are distributed through open calls on the internet. Crowdsourcing leverages human intelligence, augmented with machine learning, to produce software solutions. However, several challenges need to be tackled. For instance, designing software requirements is a major challenge during the development process. This research investigates the design aspects of the notable crowdsourcing platform TopCoder and applies machine learning approaches to identify design factors that affect developers’ decision-making during the selection process. My approach shows that design factors can be extracted to estimate software task prices and provide actionable insights into the factors influencing task success. The results will guide future research and crowdsourcing managers in applying these approaches to extract design factors and understand their impact on the development process.