Application of Predictive Artificial Intelligence (AI) Models to Estimate the Success of Crowdfunding : Metaheuristic Feature Selection

Zéman, Zoltán and Kálmán, Botond Géza and Malatyinszki, Szilárd (2025) Application of Predictive Artificial Intelligence (AI) Models to Estimate the Success of Crowdfunding : Metaheuristic Feature Selection. JOURNAL OF INFRASTRUCTURE POLICY AND DEVELOPMENT, 8 (16). No. 7934. ISSN 2572-7923

Preview

Text 7934-40291-1-PB.pdf - Published Version Available under License Creative Commons Attribution. Download (319kB) | Preview

Abstract

Statistics indicate that only a limited number of crowdfunding (CF) projects achieve their funding goals. To address this, project developers analyse key features to estimate the likelihood of a campaign's success prior to its launch. Historically, creators faced challenges in analysing these features due to resource constraints, and previous research offered only moderate success probabilities with extensive feature lists. This study introduces a novel approach utilizing a metaheuristic-based method for optimal feature selection in the CF landscape. The self-enhanced chimp optimization algorithm (COA) is employed to evaluate subsets of features with high predictive power. Open-source data from Kickstarter and Indiegogo are used, and an AI-based Convolutional Neural Network (CNN) is trained to predict the success of CF campaigns. The model demonstrates high accuracy in forecasting campaign outcomes.

Actions (login required)

Edit Item