Application of Machine Learning Techniques for Predicting Citizen Usage of Electronic Government Services

Abstract

This study examines Mexican citizens' use of e-government services through advanced machine learning models. Based on data collected from the ENDUTIH 2022 and focusing on a sample representing 30% of the total, sociodemographic variables such as sex, social strata, age, and socioeconomic level were investigated, along with technological skills and internet access patterns. The algorithms applied include K Nearest Neighbor, Support Vector Machine, Random Forest, and XGboost, with Random Forest and XGboost standing out for their precision and sensitivity. The results show that the factors studied are significant predictors of user behavior in the context of e-government, suggesting that the government can improve strategies for implementing government digital services based on these findings. However, the study acknowledges limitations, such as its focus on data from Mexican users, and recommends further research to expand the range of variables and contexts analyzed.