EEB225H1 Lecture Notes - Lecture 18: General Linear Model, Biostatistics, Dependent And Independent Variables
18 Dec 2019
School
Department
Course
Professor
Document Summary
Like a paired design, but for more than 1 treatment plus a blocking variable. Every treatment is replicated exactly once within each block. General linear model with multiple continuous predictor variables. Extension of linear regression where we measure multiple continuous predictors. Enable us to assess relationship between response variable and each of the predictors adjusting for remaining predictors. Conceptually similar to linear regression but computationally complex, estimation of parameters best represented with matrix algebra. Y = beta0 + beta1*x1 + beta2*x2 + beta3*x3 . betan*xn + eta. Beta0 = sampe as alpha, a constant. With interaction between predictor 1 and 2 e. g. beta4*x1*x2. E. g. beta1 = slope for x1 on y holding other x"s constant. In regular linear regression, model represented by a line. In multiple linear regression, model represented by a plane/hyperplane. Relationship between x and y is linear.
