Toward Project Success Forecasting in Reward-based Crowdfunding through Wide-and-Deep Computational Learning

While reward-based crowdfunding platforms centered like Kickstarter in the United States, allow entrepreneurs or well-off investors like industrial owners to gather assets and seek patron support, they have evolved into an option for generating funds via an internet platform. Conversely, these crowdfunding initiatives pose innovation challenges due to the uncertainty of receiving support and success. Both project developers and supporters may benefit from it for better knowledge and more accurate forecasting of project success. Analyzing a completed project information, this study aims to identify critical success characteristics of crowdfunding. It investigates the effectiveness of wide-and-deep learning in predicting models. These findings can help fundraisers understand the likelihood of successful fundraising initiatives.