Full Model	Ridge Regression	Ridge Regression
(Intercept)	116.3	119.3	158.5
Age	1.1	1.0	0.3
Asthma	0.7	0.5	0.4
CAD	17.0	16.9	10.0
CHF	8.0	8.6	9.4
CKD	19.5	19.8	12.8
COPD	-5.4	-4.0	4.5
Depression	31.1	29.9	10.3
Diabetes	11.0	10.9	6.9
ESRD	31.1	29.3	11.3
Female	-2.5	-2.1	0.2
HIV	70.7	68.2	23.2
Hyperlipidemia	19.9	19.5	10.2
Hypertension	-8.2	-5.5	5.8
Injury	-23.6	-22.2	-5.9
Obesity	8.9	8.7	6.0
Substance_Abuse	40.2	38.2	10.5
Wait_Time	0.7	0.7	0.2

Full Model Ridge Regression Amin Ridge Regression A158 (Intercept) 116.3 119.3 158.5 Age 1.1 1.0 0.3 Asthma 0.7 0.5 0.4 CAD 17.0 16.9 10.0 CHF 8.0 8.6 9.4 CKD 19.5 19.8 12.8 COPD -5.4 -4.0 4.5 Depression 31.1 29.9 10.3 Diabetes 11.0 10.9 6.9 ESRD 31.1 29.3 11.3 Female -2.5 -2.1 0.2 HIV 70.7 68.2 23.2 Hyperlipidemia 19.9 19.5 10.2 Hypertension -8.2 -5.5 5.8 Injury -23.6 -22.2 5.9 Obesity 8.9 8.7 6.0 Substance_Abuse 40.2 38.2 10.5 Wait_Time 0.7 0.7 0.2 In the graph above, the number of non-zero coefficient estimates associated with each value of log(1) is shown above the top horizontal axis. Question 4 Which value of 1 is larger, cuslambda. Ise or cv$ lambda. min? Which value of 1 returns the full model? Compare the coefficient estimates of Injury in the full model, the ridge regression model when 1 = Amin and the ridge regression model when 1 = Aise. As we increase 1, do the absolute values of the coefficient estimates generally decrease, remain unchanged, or increase? Are any coefficient estimates set to zero? Consider only the coefficient estimates for the predictors, not the intercept term. Answer 4