Predicting poverty index using deep learning on remote sensing and household data

Citations of this article
Mendeley users who have this article in their library.
Get full text


The main challenge for approving and implementing policies aiming at sustainable development of country is correct prediction of socioeconomic condition. Deep learning algorithms in recent researches have been identified as potential resource to be applied in this domain. Another challenge is availability of sufficient amount of data which is solved using transfer learning in Convolutional Neural Network (CNN). We used pre-trained Inception Net-v3 and Ridge regression model to estimate poverty level using publicly available dataset comprising of daylight images, nightlight images and survey data. Each cluster of samples contains households between 1-28. Its mean is 21.09, median 21 and a standard deviation is 1.36. Proposed deep learning inspired model estimates wealth-score for 28393 clusters with an r value i.e. Pearson Correlation Coefficient of 0.73, signifying r2 value i.e. Coefficient of determination of 0.54. It shows that daytime satellite images, nightlight intensity and demographic data available can be utilized for precise evaluations about the spatial scattering of monetary prosperity crosswise over different nations.




Agarwal, P., Garg, N., & Singh, P. (2019). Predicting poverty index using deep learning on remote sensing and household data. International Journal of Recent Technology and Engineering, 8(3), 164–168.

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free