An exploratory study of high-educated poverty through machine learning approach : a case study of East Java, Indonesia

Purpose - the purpose of the article is to compare the impotance factors for high-ed- ucated poverty modeling in the four cultural regions of East Java. Research methodology - using data from the Indonesian National Survey, this study employs Random Forest (RF), Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN), and K-Nearest Neighbor (KNN) algorithms as classification methods in machine learning. Findings - the analysis results show that all agorithms has high accuracy with XGboost as the best performing model. The number of household members and the use of internet technology are among the most important variables affecting high-educated poverty in each cultural region in East Java. Research limitations - the research was exclusively carried out in East Java, Indonesia, potentially restricting its applicability to other geographical areas or nations. Additionally, it relied on cross-sectional data, implying that a causal relationship between the independent and depend- ent variables cannot be inferred. Practical implications - based on the results of this analysis, it is recommended that the government and related parties focus more on improving access to education and the internet in cultural areas in East Java. Originality/Value - this study addresses a significant knowledge gap by exploring regional variations in the factors affecting high-educated poverty, emphasizing the importance of tailored strategies for each cultural region in East Java.