Bayesian Model Selection for Human Development Indicators

In this paper, we explore the use of Bayesian estimation techniques towards modeling indicators related to Human Development, in particular Gender-related measures. The main emphasis and contribution of this work is demonstrating the use of Bayesian approaches towards modeling development indicators, rather than interpreting the validity of specific models obtained from the current limited dataset. Future work will seek to acquire and utilize more comprehensive data for a greater number of indicators. There is a need to explore methods that model indicators with missing data, which is quite often the case with a majority of development indicators today. As the topic of human development is an important concern for many, this paper has been written for a broader audience. Hence, the domain of human development indicators, concepts of Bayesian estimation and statistical approaches for model selection are clearly explained.