A Machine Learning Framework for Startup Investor Matchmaking to Enhance Sustainable Innovation

Abstract

Accelerating sustainable innovation is a critical mandate for emerging economies. However, the high failure rate of startups, primarily driven by inefficient founder-investor alignment and information asymmetry, remains a systemic obstacle. In Indonesia, this challenge is further exacerbated by traditional matchmaking processes that rely on social networks, often excluding high-potential Micro, Small, and Medium Enterprises (MSMEs). This study introduces and validates Artificial Intelligence Machine Learning Excellence Economy (AIMEE), a novel hybrid matchmaking framework designed to democratize access to capital and accelerate the twin transition of digitalization and sustainability. Utilizing a pilot dataset of 94 startup-investor pairs from five provinces in Java, we developed a policy-aware algorithm that integrates financial metrics with five entrepreneurial dimensions, including Sovereignty Scores based on alignment with the Sustainable Development Goals (SDGs). The framework employs a lightweight neural network optimized with SMOTE (Synthetic Minority Over-sampling Technique) to address data scarcity and class imbalance. Empirical validation shows that AIMEE achieves an F1-Score of 0.83, Recall of 0.82, and ROC-AUC of 0.89, significantly outperforming traditional Random Forest baselines by 12%. The model effectively identifies patterns of potential collaboration, overcoming the limitations of rigid networking. These findings confirm that AI-driven matchmaking can serve as a robust digital infrastructure to operationalize the Asta Cita vision and SDG 9, contributing to a more inclusive, resilient, and competitive innovation ecosystem in Indonesia.

Keywords

• Startup
• Machine Learning
• AIMEE
• Sustainable Innovation
• Artificial Intelligence