Poverty Prediction System on Indonesian Population using Deep Learning

Abstract

The poverty rate is an important metric to measure the welfare of the population of a country. Furthermore, the yearly data of this rate is also useful for predicting the number of poor people in subsequent years. Accurate poverty prediction and measurement is an important instrument for policymakers such as the government in order to pay attention to the living conditions of the poor. In this research, deep learning was used to predict the number of poor people living in urban areas, rural areas, and both combined. The model resulting from the training process to predict the number of poor people living in urban areas, rural areas, and both combined showed successful prediction percentages of 99.9542%, 95.8261%, and 99.9419%, respectively, while the testing results for these areas were 95.0139%, 91.1321%, and 96.5633%, respectively.