Machine Learning for Fundraising Network Development in Indonesian Educational and Social Foundations

Abstract

This study explores the application of machine learning (ML) in developing fundraising networks for educational and social foundations in Indonesia, using Yayasan Pendidikan Sosial Dan Dakwah Ulul Albab as a case study. Operating in Malang City and Ponorogo Regency with eight programs requiring approximately IDR 815 million annually, the foundation faces persistent fundraising challenges. Employing a mixed-method approach, we developed an ML-based donor prediction model using Random Forest, Gradient Boosting, and Neural Network algorithms, simulated with synthetic data (n=5,000) representing Indonesian donor characteristics. Results demonstrate 87.3% accuracy in donor propensity prediction and 82.6% in donation amount forecasting. Qualitative analysis through stakeholder interviews (n=15) revealed implementation barriers including digital literacy gaps and data infrastructure limitations. The proposed ML-Integrated Fundraising Framework (ML-IFF) combines predictive analytics with culturally adapted engagement strategies, projecting 45-60% improvement in fundraising efficiency. This research contributes a contextual ML application framework for Indonesian nonprofit organizations, addressing the intersection of technological innovation and social sector sustainability in emerging markets.