Unlocking Potential: Predicting Student Entrepreneurship Performance Through Small Dataset Modelling

The role of colleges as educational institutions and knowledge centers significantly influences young people's entrepreneurship development. One of the college's efforts is a seminar on creative entrepreneurship economics. However, the output of student performance evaluations after programs is often controlled by traditional analysis because of the limited data size. This study aims to develop a performance prediction model using machine learning techniques with limited data. Model development methods used the PyCaret library to optimize data exploration, preprocessing, feature selection, model selection, training, and evaluation. Research shows that Decision Tree gives the best results, with an accuracy of around 98.33%, effectively identifying student's performance in various seminar topics. In addition, data analysis reveals correlations between variables and highlights performance determinants. This research contributed significantly to developing prediction models of student performance in entrepreneurship programs as a basis for developing future agendas.