HBM 597E - Special Topic.in Comp.Sci.&Eng

Course Objectives

1. To provide sufficient mathematical and programming background to understand the theory and application of generalized linear models.
2. To teach model selection using Akaike information criterion and parametric bootstrap.
3. To teach formal inference from more than one model.
4. To improve the student’s ability of designing, fitting and selecting generalized linear models and using them for statistical inference especially in the MATLAB programming environment.

Course Description

Likelihood function and its properties. Maximum likelihood estimation. Akaike information criterion and Kullback-Leibler divergence. Simple and multiple linear regression. Residuals, Normality, heteroscedasticity, linearity, multicollinearity. Generalized linear models (GLM). Binomial GLM and selection of the link function. Poisson and Negative Binomial GLM. Gamma and Inverse Gaussian GLM. Parametric bootstrap. Formal inference from more than one model. Model averaging. Modeling point processes using the GLM. Time-rescaling theorem.