The mathematics behind two statistical methods used for prediction are developed: principal component analysis and naive Bayesian classifier. Principal component analysis uses variables from a data set to create a smaller number of new, uncorrelated variables. The naive Bayesian classifier method predicts which class each sample in a data set belongs to based on an independence assumption, conditional probabilities, and Bayes' theorem. To illustrate each method we implement to predict first year retention for College of Wooster students. The data from College of Wooster students is used to show how the methods are implemented and how results from each method are analyzed.
Angelo, John, "Predictive Analytics on First Year Retention at the College of Wooster" (2013). Senior Independent Study Theses. Paper 960.
Bachelor of Arts
Senior Independent Study Thesis
© Copyright 2013 John Angelo