Abstract: This project aims to investigate the use of deep learning techniques, specifically Recurrent Neural Networks (RNNs), to forecast the socioeconomic deprivation status of an area utilizing orbital photography. The hypothesis is that by leveraging the information captured during the hours of daylight and the nocturnal period as observed in satellite imagery, it is possible to reliably estimate the affluence indicator of a metropolis.
The methodology involves training RNNs to learn the complex relationships between satellite imagery and the wealth index. By analyzing the visual features and patterns in the images, the RNN models are expected to capture important indicators of poverty and wealth. The models are trained using a large-scaledataset and evaluated based on their predictive accuracy. The results aims to deliver comprehensive insights into the efficacy of deep learning methodologies. andsatellite imagery for poverty prediction
Key words: Deep learning, poverty prediction, satellite imagery, Recurrent Neural Networks (RNNs), wealth index
|
DOI:
10.17148/IJARCCE.2024.13913
