Discriminatory Analysis For The Classification of Urban Poverty Levels by Province in Indonesia 2023

Abstrak

Poverty remains a significant challenge in Indonesia, particularly in urban areas that face substantial economic and social pressures. The poverty line, which reflects the minimum needs for a decent life, is used as a benchmark to determine poverty status. To measure poverty, three main indicators are commonly employed: the poverty rate (????0), poverty depth (????1), and poverty severity (????2). These indicators are interrelated and provide insight into community welfare. To better understand the factors influencing urban poverty across provinces in Indonesia, a discriminant analysis was conducted using the poverty depth index, poverty severity index, and Gini ratio as the primary variables. The analysis aimed to classify urban poverty levels and identify the most significant factors affecting them. The results indicate that the data on urban poverty rates by province in 2023 met the assumptions of multivariate normal distribution, homogeneity, and independence. The analysis revealed significant differences in the poverty depth and severity indices among the groups, while the Gini ratio did not show substantial differences. A single discriminant function was sufficient to separate the groups, with a canonical correlation of 0.729, explaining 67.89% of the data variability. Among the indicators, the Poverty Depth Index emerged as the most significant variable, while the Gini Ratio contributed the least. The centroid values for the low (-0.429) and high (2.486) poverty groups demonstrate the discriminant function's effectiveness in distinguishing between groups. The model achieved an overall classification accuracy of 91.2%, highlighting its strong performance in categorizing provinces based on urban poverty status.