CLASS EXAMPLE FOR MODEL VALIDATION AND SELECTION
Example: Surgical Unit. A hospital surgical unit wanted to predict survival in patient undergoing a liver
operation. A random sample of 108 patients were available for analysis. From each patient record, th
CLASS EXAMPLE FOR MEASURES OF INFLUENCE AND OUTLIERS
Example: Body Fat. Response =body fat, we will use 2 predictors: triceps thickness X1, and thigh
circumference x2.
Regression Analysis: bodyfat versus tricepsX1, thighX2
The regression equation is
bodyf
MULTICOLLINEARITY, ITS EFFECTS AND DIAGNOSTICS (7.6 and 9.5)
How does the correlation of explanatory variables affect regression models?
Example: (uncorrelated explanatory variables). Work crew productivity (Y) is studied as response to the
crew size (X1)
BODY FAT LECTURE EXAMPLE INTERACTION MODELS (8.2)
Body fat of 20 people was recorded along with the circumferences of their triceps (X1) thigh (X2), and midarm
(X3). We want to model the body fat as a linear function of these predictors AND THE INTERACTIO
MULTWLE REM-1.36900 I
SLRM one 'PREDICXOR 09w.qu
y = 13.. 1- gm + a one vmmoe. newt.
77 1 MON 39 m swim,
mfomse pwétukoz. 10 BE USEFUL.
EXAMPLE -. Y = sum»! , X = 739:!!!» 0:- cinnamon
QURH MN 36 To. QMHE . wuswm
a-DDIHONGL Pneumwns 1'0 IMPMVE THE MOD
STAT 757 QUALITATIVE AND QUANTITATIVE VARIABLES IN REGRESSION
DUMMY VARIABLES
EXA L PLE: Innovation in insurance. Relate the speed with which a particular insurance innovation is adopted (Y)
to the size of the insurance rm (X1) and the type of the rm. X