Ranking algorithms are utilized for a variety of purposes. Google utilizes a page rank algorithm to hone its search engine and recommender systems utilized by Amazon, Netflix and Youtube all use rank algorithms to tailor their suggestions to each individual user. While a ranking algorithm is simply an attempt to order a set of objects along some criteria, there are a variety of ranking algorithms that give different results for different inputs. This project focuses on a technique known as the Keener Method, a linear algebra technique, to make predictions for 'March Madness', the annual men's college basketball tournament. The Keener Method will be used to rank teams along different metrics of performance. These rankings will then be aggregated into a single vector. The accuracy of this model will be evaluated by simulating this algorithm on past data. This project will also attempt to find the optimum metrics of performance for prediction, which may improve the Keener Method for future applications.
