Monday, May 8, 2017

URSP Student Osaze Shears Explores How to Create and Filter Multivariate Data Sets

My name is Osaze Shears and I am a junior Computer Engineering major conducting research in the fields of machine learning and computer hardware acceleration. Since high school I have always had a passion for conceptualizing the various applications of general artificial intelligence. It amazed me to see the research being conducted in the field, and applied to robotics and data analytics. Early in my junior year I started to learn more about various machine-learning algorithms as a means to predict dynamic real world information.

My research career in machine learning began when I met with my faculty advisor in the Department of Electrical and Computer Engineering, Houman Homayoun, and he gave me the opportunity to conduct research with his graduate student, Maryam Heidari. We realized that the real estate market has heavy implications for investors and prospective home owners when it comes to making decisions regarding buying and renting properties. The price value of the house is among the most important features considered in these decisions, and so it is useful to gain accurate approximations of a house’s value in order to determine where it ranks in comparison to similar properties.

Because I was interested in learning about the different ways in which machine learning could be applied, this research project has allowed me to explore how to create and filter multivariate data sets in order to improve the prediction capabilities of machine learning algorithms. By going through this process I have gained insight on what house features carry the most weight when the value of a house is determined, and furthermore how to decide which modeling algorithms are best to implement when dealing with various data sets. This research will definitely assist me in my future career goals when I am creating computer hardware devices that are used to accelerate the rate at which these computations can be trained and applied.