A Robust Predictive Modelling of Nigeria’s Population Growth Rate Using Partial Least Square Regression

Abstract

Nigeria, a developing nation is experiencing the overwhelming effects of her exponentially ever-increasing population. The resultant effects are clearly evident for all stakeholders to see and feel. Researches have been carried out to study, explain and recommend solutions to this lurking epidemic. But unfortunately, numerous researchers have failed to address key issues in regression modelling as used in their studies, some of such issues are; using Wald’s statistic as a variable selection tool rather than the much consensus purposeful variable selection techniques, ignoring the existence of multicollinearity and also missing data. These issues are enough to render the findings in most studies reviewed inadequate, invalid and misleading to be used as a policy-making tool. In this study, the aim is to build a robust predictive model of the Nigeria population growth rate taking into account the aforementioned issues in regression modelling hitherto ignored by some researchers who had used almost this same variables used in this current study. As it would have been expected, death rate, maternal deaths and infant deaths all had negative signs indicating an opposing relationship between these variables and Nigeria population growth rate. The assessment carried out showed that our model has high predictive power, hence, could be used to predict future Nigeria’s population growth rate.