Modeling the Determinants of Entrepreneurial Success and Failure in Newly Created Moroccan SMEs: A Machine Learning Approach

Abstract

In Morocco, SMEs play an undeniable role in economic development and job creation to combat unemployment. However, studies of newly created businesses, mainly SMEs, reveal a very high mortality rate. This article aims, firstly, to identify the endogenous and exogenous factors that explain the success and/or failure of entrepreneurship in newly created SMEs; secondly, to develop a predictive model based on Machine Learning techniques. Thus, statistical tools such as the chi-2 test, contingency coefficient, correlation matrix, and principal component analysis (PCA) were used to analyze the data and test the hypotheses. Next, binary logistic regression enabled us to model the relationship between the independent variables and the dependent variable, while measuring the impact of each explanatory variable. Finally, Machine Learning techniques were applied to identify the most significant variables in our conceptual model. These variables will be integrated into our predictive model based on the Random Forest technique. The results show that out of the 27 variables comprising our conceptual model, only 12 variables have a significant influence in explaining the entrepreneurial situation of entrepreneurs in newly created SMEs, with a dominance of factors aligned with the resource-based and skills-based approach.