Machine Learning Valuation of Quebec Housing Market

Most academic literature and practices in the real estate industry use traditional valuation models to predict house prices. While machine learning models have been used more heavily in the finance literature, it is less applied among real estate researchers. While traditional property valuation models rely on simple relationships between the price of a property and each property characteristic, machine learning models allow for complex relationships and can solve such relationships. Using Quebec housing data from real estate brokers, we will predict house prices using machine learning models and aim to generate better predictions of house prices. We will also develop a benchmark for renovation in houses to improve our predictions. Additionally, we will use neighborhood data and house images to better predict house prices.

Faculty Supervisor:

Erkan Yönder

Student:

Mehdi Rasteh;Alexander Sarandiev;Elaheh Nikbakht

Partner:

Squarefeet.ai

Discipline:

Finance

Sector:

Professional, scientific and technical services

University:

Concordia University