Forecasting Profitability of Real Estate Assets using Machine Learning

This research project aims at applying machine learning over the existing financial forecasting methods currently employed in the commercial real estate industry. Businesses are actively collecting more data than what can be analyzed effectively using the standard spreadsheet models which have become industry standard over the past few decades. Machine learning algorithms are known to be able to extract complex relationship between many variables in data which make them perfect for an application geared towards forecasting the financial performance of commercial real estate assets. This task involves aggregating large amounts of data specific to the asset in question such as revenue, expense and leasing information as well as relevant economic data including rent growth and employment figures. This project will evaluate the performance of some of the most common machine learning algorithms and their relevance to predicting the performance of commercial real estate. Much of the research that currently exists at the intersection of finance and machine learning revolves around the public financial markets or the mass appraisal of residential real estate.

Faculty Supervisor:

Elkafi Hassini;Kai Huang

Student:

Andrew Foresi

Partner:

One Cornerstone Solutions Corp

Discipline:

Computer science

Sector:

University:

Program:

Accelerate