The massive data collected by higher education institutions (HEIs) provides these institutions with opportunities to gain insights from the data so as to make sound strategic and operational decisions. However, some of the HEIs experience difficulties in identifying suitable Machine Learning (ML) algorithms to analyze their valuable data. At the same time, many HEIs are struggling to improve student performance and success. Several existing tools that are aimed at predicting student performance fail because they either provide such prediction late in student years of study or they consider factors, which are not in the control of students (eg. Demographic factors). To assist in addressing the abovementioned challenges, the study set out to find suitable supervised data mining algorithm(s), which would assist to predict the likelihood of a student engaging in a studying activity. By finding a suitable technique to address this problem, HEIs would not only be aided in improving student success, but they will also be assisted in identifying a suitable ML technique to analyze their data. To meet this study objective, a publicly available dataset on students’ performance with 145 entries was obtained from Kaggle. Logistic Regression, Random Forest, Neural Networks and AdaBoost were the four ML algorithms considered in this study. When all the input features were considered, the Neural Networks outperformed the other three with an average accuracy of 64%. However, when the important features were considered, both Neural Networks, and AdaBoost models were found to be best-performing models at 70% accuracy level.