Analyzing Social Expenditure
for a Sustainable Society using Machine Learning: Linear Regression
Young Kim
Korea International School
Abstract
Social expenditure, which encompasses cash benefits, direct in-kind provision of goods and services, and tax breaks with social purposes, is an important source of support for disadvantaged or vulnerable groups, such as low-income households, the elderly, the disabled, the sick, the unemployed, and the young. Due to its importance in the socio-economic sphere, it is essential to discover measures for optimizing a country’s net social expenditure. As one of numerous data mining techniques, linear regression is an analysis methodology that provides a synthesis of inputs to calculate an output variable. The Corruption Perception Index (CPI), the COVID-19 Case Fatality Rate (CFR), the Gross Domestic Product (GDP), and the Environmental Performance Index (EPI) are the four factors we synthesize for this study. The results indicated that the ideal multilinear regression model for forecasting a country’s social expenditure was the combination of all four parameters. In addition, the data suggest that EPI had the highest association with social expenditure, but GDP, which had previously been the primary factor in determining a country’s net social expenditure, had the lowest correlation among the four variables examined. In the future, we intend to combine other factors and different prediction techniques, such as feed-forward neural networks, to construct a more accurate prediction model.
