Agricultural Economics Research Review

The first agenda in the Sustainable Development Goals outlined by UNDP is to eradicate poverty in its all forms and from everywhere. The nations across the world have initiated several programmes towards achieving this goal. However, a major problem relates to the availability of data on poverty, especially in the developing countries. This is a limiting factor in targeting poverty and assess the impacts of poverty reducing programs. In this paper, we explore the use of satellite night light data and machine learning algorithms (Artificial Neural Network) to predict rural poverty at sub-national level in India. The statistical superiority of the prediction has been tested against multiple imputation models. The results show that machine learning technique has a better predictive power, compared to other imputation models. These predictions can be gainfully used to complement the existing datasets for tracking and targeting poverty and assessing the impact of poverty-eradication programmes.