An Optimized Machine Learning Model for Population Growth Prediction Using Artificial Neural Network and Genetic (Neuro-Genetic) Algorithm

Abstract

With the global population growth reaching over seven billion persons on the
planet, issues of desertification, famine, global warming and climate change are
presently in the front burner of global discourse as the need for proper and more
accurate population prediction for economic and national development become
expedient. This paper therefore proposes’ the optimization of machine learning
model for population prediction in Nigeria using artificial neural network and
genetic algorithm. The model was adopted to dataset for the national population
commission (NPC) containing annual population growth from 1950 – 2021. The
Agile Software Development Methodology (SDM) was applied to real-life data to
evaluate the efficiency of the model. An iterative approach was adopted to time
series data in other to examine the applicability of the proposed model. The model
was implemented in Java Apache NetBeans using 70% of the dataset for training
and 30% as test data. The model yielded an accuracy of 76% with a Root mean
squared errors (RMSE) of 8.21%, mean Absolute percentage errors (MAPE) of
6.4%. Mean squared errors (MSE) of 5.67%, MEA of 23.29% and MAD of 49.23%
when compared to existing models in literature.