Technology-mediated method for prediction of global government investment in education toward sustainable development and aid using machine learning and classification Academic Article in Scopus uri icon

abstract

  • Predicting and monitoring of global government investment, e.g., using AI-based method, is becoming an emerging topic aimed at applying technological-based solutions to address critical issues for benefit of human. Using data from UNESCO's Institute for Statistics (UIS) about government expenditure in education on sustainable development (SDG) between 1970-2020; this study implements a machine learning method for prediction of the values or rate of governments' expenditure as a proportion of gross domestic product (GDP%), done by training and testing key features (year of investment and regions' classification) that are technically considered adequate for prediction of the investment values by region. The proposed method was designed based on the cross industry standard process for data mining (CRISP-DM) methodology, and implemented using supervised machine learning technique such as K-Nearest Neighbor (KNN), and validated using k-fold cross-validation. The results prove that performance of the executed model was efficient for prediction of values of the global government expenditure in education with Precision=0.77, Recall=1.00, Accuracy=0.78, F1-score=0.87, and low Error-rate of 0.22, respectively. Also, the study empirically shed light on both the socio-technical and pedagogical implications of the results and output towards achieving a sustainable educational practice, particularly SDG4 that promotes quality of education, and decision/policy making by the governments, educators, and policy makers. © 2023 IEEE.

publication date

  • January 1, 2023