Chapter02

Chapter02 - STAT 563 Regression Methods Simple Linear...

• Notes
• 99

This preview shows pages 1–11. Sign up to view the full content.

STAT 563: Regression Methods Simple Linear Regression

This preview has intentionally blurred sections. Sign up to view the full version.

What is a Regression Model? Essentially a mathematical formula relating one variable, referred to as a response variable, to one or more other variables, called predictor variables Often used in exploratory studies and generally not intended for causal relationships Uses of regression models include Data description Parameter estimation Prediction and estimation Control Example: predict average annual salary (Y) using time (t) as a predictor variable A regression model is a model of the conditional expectation E(annual salary|time)
Simple Linear Regression Suppose we have n pairs (Y i , X i ) of data on (salary, time) Specify the model for (X 1 ,X 2 ,…X n ) Normality is not needed for estimation but required for inference Fit the model: estimate ( β 0 1 29? Least Squares Computation starts with finding Inference: Draw conclusions on β 1 based on n observations ) , 0 ( ~ 2 1 0 nxn i i i I N X Y σ ε ε β β + + = ( 29 2 1 1 0 , 1 0 1 0 min arg ) ˆ , ˆ ( = - - = n i i i X y β β β β β β ) ˆ , ˆ ( 1 0 β β

This preview has intentionally blurred sections. Sign up to view the full version.

Least Squares Least squares regression finds the line that minimizes Normal Equations: = - - = n i i i X y S 1 2 1 0 1 0 ) ( ) , ( β β β β 0 ) ( 2 0 ) ( 2 1 1 0 1 1 1 0 0 = - - - = = - - - = = = n i i i i n i i i X X y S X y S β β β β β β
Least Squares Simplifying and solving for (substituting the estimates for the parameters): The fitted model is: ) ˆ , ˆ ( 1 0 β β = = = = = = - = - - = = - = n i i n i i n i i xx n i i i yx xx yx X n X y n y X X S X X y y S where S S X y 1 1 1 2 1 1 1 0 1 , 1 , ) ( ) )( ( ˆ ˆ ˆ β β β X y 1 0 ˆ ˆ ˆ β β + =

This preview has intentionally blurred sections. Sign up to view the full version.

Geometry of Least Squares The least squares estimate of the data is an orthogonal projection of data vector onto the independent variable subspace
Geometry of Least Squares For each pair ( β 0 1 29 the vector with components The S can now be expressed as Minimizing this length over ( β 0 1 29 finds the vector closest to the plane L spanned by 1 , X . Or, the least squares projects the vector y onto the plane L. ) , ( 1 0 β β P i i X P 1 0 ) , ( , 1 0 β β β β + = 2 ) , ( 2 1 ) , ( , 1 0 1 0 ) ( β β β β P y P y n i i i - = - = ) ˆ , ˆ ( 1 0 β β P

This preview has intentionally blurred sections. Sign up to view the full version.

Residuals The residuals e i are defined as In vector notation , or, e is the projection of y onto the orthogonal complement L of the plane L spanned by 1 , X . This implies The vector of residuals e is independent of i i i y y e ˆ - = y y e ˆ - = 0 ˆ 0 0 1 1 1 1 = = = = = = = = = e y y e e X X e e e T i n i i T n i i i T n i i y ˆ
Estimating σ 2 Required to Test hypotheses Construct confidence intervals Ideally, we would like to have an estimate independent of adequacy of fitted model Only possible if we have multiple observations on y for at least one value of X or if prior information is available on σ 2 In absence, we will use error (or residual) sum of squares

This preview has intentionally blurred sections. Sign up to view the full version.

Estimating σ 2 If ( β 29 is known then and Now, an unbiased estimate can be obtained from is an unbiased estimate of σ 2 i i i X y 1 0 β β ε - - = 2 2 1 0 1 2 2 ) , ( n n i i SSE χ σ β β ε ε = = = = 2 1 2 1 σ ε = = n i i n E 2 ε
This is the end of the preview. Sign up to access the rest of the document.
• Spring '07
• Unknown

{[ snackBarMessage ]}

What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern