STAT 425 - Fall 2024 - Sections 1 and 2
STAT 425 - Fall 2024 - Sections 1 and 2
-
From Douglas Simpson
Considers the extension of ANOVA models to include factors that are themselves random variables. This type of modeling is useful to account for batch correlations in the… -
From Douglas Simpson
Discusses how blocking is used in experimental design, and how the design affects how we analyze the data. In randomized complete blocks designs, the treatment… -
From Douglas Simpson
Discusses the difference between observational studies and experimental studies. Introduces key ideas and examples of experimental design including randomization,… -
From Douglas Simpson
Discusses analysis and interpretation of results for unbalanced two-way designs. In this case we rely on the general linear model formulation and test each effect after… -
From Douglas Simpson
Extends the analysis of variance to models with two or more factor variables. By doing two-way and multiway analysis of the factor effects we can determine whether there… -
From Douglas Simpson
Suppose we detect differences between the group means in the analysis of variance. We would like to know how those means differ. Characterizing the mean differences… -
From Douglas Simpson
A one-way anova design is a comparative study in which we compare responses between several treatments, groups or factor levels. Often the subjects or experimental units… -
From Douglas Simpson
Examples and computational details for Lasso regression, i.e., L1 penalized least squares. -
-
From Douglas Simpson
Introduces penalized least squares methodology for stabilizing linear regression with correlated or high dimensional predictors. Two key methods are Ridge Regression,… -
From Douglas Simpson
Discusses principal component regression as a way to both understand the variation in the regression predictor variables, and potentially reduce the dimension of the… -
From Douglas Simpson
We walk through R code for variable selection in multiple linear regression, focusing on analysis of state level data relating average life expectancy with key… -
From Douglas Simpson
In selecting predictor variables to include in a regression model, an important consideration is the bias-variance tradeoff: adding variables to the model generally… -
From Douglas Simpson
Analysis of covariance (Ancova) models include both numerical and categorical predictor variables. We discuss building and interpreting Ancova models, emphasizing the… -
From Douglas Simpson
Includes computational details for B-spline and natural cubic spline curve fitting. Also demonstrates how to perform cross-validation for selecting the optimal number… -
From Douglas Simpson
Regression splines provide a flexible family of curves that can be fit to data using ordinary least squares linear model methods, yet closely model arbitrary smooth… -
From Douglas Simpson
Often the expected response is nonlinear as a function of a predictor variable. These notes show how we can fit polynomial curves to such nonlinear response data while… -
From Douglas Simpson
Provides computing details for examples of generalized least squares regression and lack of fit testing. The coding of factor variables for the lack of fit test is… -
From Douglas Simpson
Considers how to test whether the systematic part of a linear model fits the data. The key idea is to compare the model-based estimate of the error variance to either… -
From Douglas Simpson
If the errors in the linear model are heteroscedastic or correlated with known structure, or in ways that can be modeled, generalized least squares (GLS) estimation can… -
From Douglas Simpson
Demonstrates the computational details for collinearity diagnostics in analysis of the car seat position data, and shows how the model can be improved by selecting a… -
From Douglas Simpson
Collinearity between explanatory variables can make it difficult to interpret the individual effects of variables in the model. Building on the added variable / partial… -
From Douglas Simpson
Methods for detecting and characterizing unusual observations in the data relative to the regression model. Unusual observations include: 1) high leverage observations;… -
From Douglas Simpson
R Markdown presentation of several examples for graphical and formal statistical evaluations of modeling assumptions. Provides additional details for several examples in… -
From Douglas Simpson
Discusses how to check for violations of the Gauss-Markov modeling assumptions of constant variance, uncorrelated errors and linearity of the mean function. We consider… -
From Douglas Simpson
Demonstrate coding and interpretation of diagnostic regression plots and statistics for several data examples. Includes pairwise scatter plots, standardized and… -
From Douglas Simpson
Demonstrates coding of Anova F tests, permutation F tests, confidence intervals, prediction intervals and the construction of elliptical confidence regions for two… -
From Douglas Simpson
This video discusses both individual and simultaneous confidence intervals and regions for linear regression coefficients, along with model-based confidence intervals… -
From Douglas Simpson
This video covers hypothesis testing for coefficients in the linear model, based on Gaussian error assumptions. We use coefficient t-tests for individual explanatory… -
From Douglas Simpson
This video shows how to derive the expectation vector, covariance matrix and joint distribution for LS estimators in multiple linear regression. Using the same general…