Property evaluation in Real Estate using machine learning · Electronic Theses and Dissertations

Author: Jiri, Godwin S.
Title: Property evaluation in Real Estate using machine learning
Abstract: This research investigates the application of supervised machine learning techniques in the prediction of property prices, with a specific focus on residential real estate in Bulawayo, Zimbabwe. The study addresses the limitations of traditional manual valuation methods, such as human error, delays, and inconsistencies, which often lead to inaccurate pricing and affect key stakeholders including buyers, sellers, and real estate agents. To provide a data-driven solution, a machine learning-based system was developed using the Random Forest algorithm due to its robustness and high accuracy in predictive tasks. The model was trained on a real estate dataset comprising various property attributes such as location, number of bedrooms, land size, and property type. Performance evaluation metrics of the model produced the following results: A Mean Squared Error (MSE) of 4,669,713,572.27, Root Mean Squared Error (RMSE) of 68,335.30, R-squared (R²) of 0.871, Mean Absolute Error (MAE) of 43,224.97, and Mean Absolute Percentage Error (MAPE) of 0.25%. These values indicate strong predictive performance, validating the model’s effectiveness in estimating property values. The developed system allows users to input property features to obtain estimated values and also provides functionality to load your property in the market after evaluation. This study contributes to enhancing property valuation practices by introducing a reliable, automated, and user-friendly approach to real estate pricing using machine learning.
Date: June 2025
Publisher: BUSE
Keywords: machine learning
Supervisor: Mr. P Chaka

Item sets: Department of Computer Science

Media: GODWIN S JIRI.pdf