Date : 23/06/2025
Hour: 20:45
Student Name & ID: NABIL ABDUSALAM.A IBRAHIM // STUDENT ID : (245 105 403)
Supervisor: Prof. Dr. Ömer KARAL
Link or Room: https://meet.google.com/hhb-bmhu-ggb
Meeting Link (Online)
Abstract:
Crowdfunding platforms have become a prominent source of funding for startups, especially in emerging economies like Turkey. Despite the rapid adoption of crowdfunding, predicting campaign success remains a challenge. This study uses deep learning and machine learning methods, focusing on Random Forest and Logistic Regression, to identify the key factors influencing campaign outcomes. Data from 12,000 Turkish crowdfunding campaigns are analyzed, and results show that goal amount, campaign duration, and project category significantly impact success rates. Random Forest achieves an accuracy of 90% and an AUC-ROC of 94%, providing robust predictive performance.
Keywords: Machine Learning, Random Forest, Logistic Regression, Confusion Matrix, Predictive Analytics.