Abstract
Despite extensive research on AI's theoretical benefits in entrepreneurship, few studies compare machine learning models' effectiveness using real-world data or address challenges like model interpretability and overfitting. This study investigates how AI-driven big data analytics enhances entrepreneurial decision-making in the digital economy by evaluating four machine learning models-Decision Trees, Random Forest, Gradient Boosting, and Histogram-Based Gradient Boosting-to predict AI service focus. The results reveal that Gradient Boosting outperformed others with a testing R² of 0.9914, identifying company reputation and location as the most influential predictors of AI adoption. These findings challenge assumptions about organizational size's role in digitalization, emphasizing the strategic value of brand and geography. Key limitations include overfitting in Decision Trees and Random Forest, and reliance on static datasets that constrain real-time adaptability. The results demonstrate AI's potential to reduce uncertainty in entrepreneurial strategy, offering actionable insights for market entry and investment decisions. Future research should incorporate real-time data streams and hybrid AI-human frameworks to improve generalizability.