A Comparative analysis of machine learning models for housing price prediction in Nairobi Metropolitan Area

Abstract

Background: The Nairobi Metropolitan Area has undergone significant transformation in recent years, driven by urbanization and economic opportunities. As the nation’s capital and primary economic hub, it has attracted a growing population in need of housing solutions. This population influx has led to a dynamic housing market with fluctuating property prices, prompting the need for accurate housing price predictions. These predictions are crucial for various stakeholders, including homebuyers, developers, investors, and policymakers, as they guide decisions related to property investment, affordability, and policy formulation. Methodology: The study employed advanced machine learning techniques to predict housing prices in the Nairobi Metropolitan Area. Various models, including spatial random forest and spatial autoregressive models, were utilized to develop predictive frameworks. Evaluation of these models was conducted using key metrics such as coefficient of determination (R-squared), root mean square error (RMSE) and mean absolute error (MAE). Results: The results of the study revealed that the spatial random forest model emerged as the most effective tool for predicting housing prices in the Nairobi Metropolitan Area. The analysis also identified several key features that significantly influenced property values, including property size, number of bedrooms, and type of area. These findings underscored the importance of both structural attributes and neighbourhood characteristics in determining house prices in the dynamic urban environment of Nairobi. Conclusion: The study underscores the critical role of accurate housing price predictions in guiding decision-making for various stakeholders in the Nairobi Metropolitan Area. Leveraging advanced machine learning techniques, particularly the spatial random forest model, provides valuable insights into the dynamics of the local real estate market. Keywords: Housing prices, Nairobi Metropolitan Area, Machine learning, Predictive modeling, Spatial analysis