All data analysis requires some sort of apriori assumptions on the underlying model. However, the more stringent the assumption, the more likely it is to be incorrect. The main idea behind nonparametric inference is to use as few assumptions as possible. Typically this means that the statistical models are infinite-dimensional. This makes them much more flexible, but also more complex to analyse.
In this course, we will study several aspects of nonparametric inference including density estimation, nonparametric regression, the bootstrap, and shape-restricted inference. Students are assumed to have a solid general knowledge of parametric statistics (theory and practice), including likelihood inference and linear models. This course will also involve much computing and simulation, and prior knowledge of R is beneficial although not necessary.
Instructor: Hanna Jankowski
E-mail: hkj[at]mathstat.yorku.ca
Please include "[4230/6634]" in the subject of your e-mail.
Properly-written text messages only, no html, no "texting".
Office: N621B Ross
Office hours:
MW11-12, or by appointment. No drop-ins, please.
[Course Syllabus] [revised grading scheme]