Publication Date

Spring 5-25-2021

Degree Type

Master's Project

Degree Name

Master of Science in Computer Science (MSCS)

Department

Computer Science

First Advisor

Ching-seh Wu

Second Advisor

Robert Chun

Third Advisor

Fabio Di Troia

Keywords

Real Estate Appraisal, Machine Learning, Convolutional Neural Network, Ensemble Learning, Image Evaluation

Abstract

Real Estate Appraisal is performed to evaluate properties during a range of activities like buying, selling, mortgaging, or insuring. Traditionally, this process is done by real estate brokers who consider factors like the location of a house, its area, the number of bedrooms and bathrooms, along with other amenities to assess the property. This approach is quite subjective since different brokers may arrive at a different quote for the same property depending on their analysis. The development in machine learning algorithms has given rise to several Automated Valuation Models (AVMs) to estimate real estate prices. Real estate websites use such AVMs to provide an estimated price to potential buyers and sellers of properties. However, these models do not consider the impact that the appearance of a house has on its price.

Recent advancements in Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance for several computer vision tasks. This project uses a CNN to evaluate and score the image data associated with a house. This information was then combined with other property-related data and three ensemble learning models, namely, Random Forest, Gradient Boosting, and Extreme Gradient Boosting (XGBoost), were then trained to estimate real estate prices. The performance of these models was compared with each other and it was found that the XGBoost model achieved the best performance with a MAPE score of 9.86%. Although the image-based model performed better, the XGBoost model which did not consider image data also achieved comparable results with a MAPE score of 10.33%.

Available for download on Wednesday, May 25, 2022

Share

COinS