Using Online Data to Predict Startup Success
In recent years technology startups have both become very prevalent in society and attracted the interest of many researchers. The question of how to predict the probability of success of a startup is crucial to problems in fields such as portfolio optimization and business strategy. In previous works researchers have found that by using multiple data sources to build interaction networks it is possible to both understand and model an underlying system in a way that yields powerful predictions about the system. The goal of my SuperUROP is to use online data to build a network-based model of the startup ecosystem that is able to make probabilistic predictions about whether or not a startup will succeed in a given time period.

My interest in predictive analytics and machine learning started years ago and has since grown through multiple UROPs and classes such as 6.867 and 6.437. I want to participate in the SuperUROP program in order to both hone my research skills and have the opportunity to create a predictive model that is impactful on the world.