Predicting the Income Groups and Number of Immigrants by Using Machine Learning

Abstract

One of the most significant challenges that humanity has ever faced is migration. Urban planning, commerce, disease transmis
sion, pandemics, and public policy are just a few of the many areas that rely on precise predictions of human movement. An 
increasingly ubiquitous tool, Artificial Intelligence (AI) allows us to forecast migratory patterns. The goal of this research is to 
use ML algorithms to foretell immigration numbers and income brackets. The research included two separate applications. Two 
separate projects were undertaken with the aim of forecasting immigrant income brackets and total numbers. This research 
made use of data collected by the World Bank. Initially, the research used SVMs, NBs, LRs, and KNNs (K-Nearest Neighbors) 
in its application. The study’s second application made use of the Xgboost and Random Forest (RF) algorithms. Xgboost 
achieved a success rate of 98.37%, RF 96.42%, LR 86.04%, SVM 83.72%, KNN 83.72%, and NB 69.76% according to the 
study’s data. With the LR and Xgboost algorithms, the research found that the applications were the most successful. Generally 
speaking, the human migration prediction machine learning models used in this research will provide a versatile foundation for 
modeling human migration in various what-if scenarios.