A Comparative Study of Machine Learning Algorithms for Predicting Project Outcomes

Crowdfunding platforms are online services that allow individuals and organizations raise capital for their various projects or initiatives. These platforms transform the way people can fund their ideas, products or activities. Among these platforms, Kickstarter is one of the most well-known and successful. It has facilitated the launch of thousands of projects, aiding creators bring their ideas to life with the support of generous backers, however, determining the outcomes of the project is a crucial decision-making for both creators and funders. Machine learning algorithms offer powerful tools for analyzing large amounts of data and identifying themes and patterns that can predict whether a project will be successful, fail, be canceled, remain live, or be suspended. By employing machine learning algorithms such as K-nearest Neighbors (K-NN), Decision Trees (DT), and Neural Networks (NN), three distinct models were developed to predict the status of Kickstarter projects. The results indicated that the Decision Tree algorithm outperformed both K-NN and NN in terms of accuracy, achieving a performance rate of 89.02%. This study highlights the significant potential of machine learning algorithms to predict outcomes that are essential for decision-making for creators and backers alike in crowdfunding-related activities.