Toward predicting entrepreneurial activity among Moroccan Entrepreneurs: A machine learning approach

Abstract

Entrepreneurial activity, a subject of enduring intrigue among scholars, continues to captivate attention, especially in distinct contexts such as Morocco. This study undertakes the formidable task of comprehending and forecasting entrepreneurial activity using the comprehensive Global Entrepreneurship Monitor (GEM) dataset for Morocco. Employing a diverse range of machine learning classifiers, including logistic regression, random forest, support vector machines, gradient boosting, and K-nearest neighbors, our research excels in predicting entrepreneurial activity with remarkable accuracy. Notably, support vector machines emerge as the most potent classifier, achieving an impressive accuracy rate of 95.33%. These findings transcend the inherent complexities of understanding the infrequent nature of entrepreneurial activity, providing invaluable insights into predictive modelling within the Moroccan entrepreneurial landscape. This research not only advances our comprehension of entrepreneurship but also paves the way for informed policymaking and the nurturing of a thriving entrepreneurial ecosystem. This underscores the critical importance of effective data handling and model refinement in achieving these milestones.