Supervised machine learning predictive analytics for alumni income

Abstract

Background: This paper explores machine learning algorithms and approaches for predicting alum income to obtain insights on the strongest predictors and a ‘high’ earners’ class. Methods: It examines the alum sample data obtained from a survey from a multicampus Mexican private university. Survey results include 17,898 and 12,275 observations before and after cleaning and pre-processing, respectively. The dataset comprises income values and a large set of independent demographical attributes of former students. We conduct an in-depth analysis to determine whether the accuracy of traditional algorithms can be improved with a data science approach. Furthermore, we present insights on patterns obtained using explainable artificial intelligence techniques. Results: Results show that the machine learning models outperformed the parametric models of linear and logistic regression, in predicting alum’s current income with statistically significant results (p < 0.05) in three different tasks. Moreover, the later methods were found to be the most accurate in predicting the alum’s first income after graduation. Conclusion: We identified that age, gender, working hours per week, first income and variables related to the alum’s job position and firm contributed to explaining their current income. Findings indicated a gender wage gap, suggesting that further work is needed to enable equality.