Hybrid neural network with image caption generator for spatio-temporal income inequality and poverty estimation: A case study in Ecuador
摘要
We analyze the economic characteristics of income distribution discrepancy using Ecuador as a case study. The study discusses the difficulties and significance of accurately estimating income and poverty at the national and subnational levels in time. It emphasizes the emergence of novel tools for estimating poverty, such as machine learning and the combination of satellite pictures, economic data, and time series analysis. Our hybrid neural network model forecasts income and poverty trends by combining convolutional and recursive networks. The study highlights the importance of foreign investment, household consumption patterns, agricultural productivity, infrastructure investments, financial activity, natural resource management, public administration, and industrialization. It suggests interventions such as infrastructure investments, economic incentives, and local governance capacity building to address regional inequalities and promote sustainable economic growth.