Enhancing Rental Cost Predictions for Student Housing in Lokoja: A Comparative Analysis of Machine Learning Models

Abstract

Accurately determining rental costs for properties in specific locations is crucial for influencing market dynamics and housing decisions for both homeowners and prospective tenants. This study presents a data-driven predictive analytics model designed to forecast house rents for students in Lokoja, North-Central Nigeria, using machine learning techniques. The research involved meticulous data collection from students residing within the three major student-dominated localities of Adankolo, Felele, and Crusher in Lokoja. Data preprocessing and feature selection were undertaken to ensure the quality and relevance of the dataset. The dataset comprises eleven predictors and 300 observations. Eventually, seven predictors and 265 observations were used for modeling, with a split of 80% for model training and 20% for testing. Three machine learning algorithms - Random Forest Regression, Linear Regression, and Decision Tree Regression—were evaluated for their predictive accuracy. The Random Forest Regression model emerged as the most accurate, achieving an R2R^2R2 value of 95.2%. Correlation analysis confirmed a strong positive relationship between the selected features (e.g., number of rooms, furniture) and rental prices. A user-friendly interface was developed to facilitate rent predictions based on user inputs. The findings underscore the model’s robustness and its potential for real-world application in the real estate market. Suggestions for future work include incorporating additional features, expanding the geographic scope, and exploring advanced machine learning techniques to further enhance model accuracy and applicability.