[R] write out a complete GAM fitted model equation, using parameter estimates

[greg holly]

Hi all,

I have the following results from a study. I ran this model for binary
distributed outcome log(event/non_event). I need to write out a complete
GAM fitted model equation, using parameter estimates given below. I have
difficulty putting the parameter estimates of Spline part in the fitted
model. Your help is much appreciated.

Merry Christmas,
Oslo





Regression Model Analysis
Parameter Estimates
ParameterParameter
EstimateStandard
Errort ValuePr > |t|
Intercept -0.59214 1.47184 -0.40 0.6875
Study XX -0.53056 0.15266 -3.48 0.0005
Study YY 0 . . .
sex 1 -0.23585 0.22393 -1.05 0.2924
sex 0 0 . . .
Linear(cal_a) 0.01551 0.00718 2.16 0.0309
Linear(cal_B) -0.01627 0.00950 -1.71 0.0870
Linear(vit) 0.84962 0.17092 4.97 <.0001
Linear(mg) -2.21676 0.19107 -11.60 <.0001

Smoothing Model Analysis
Fit Summary for Smoothing Components
ComponentSmoothing
ParameterDFGCVNum
Unique
Obs
Spline(cal_a) 0.999911 3.381362 3.007482 76
Spline(cal_B) 1.000000 0.140625 10.741715 1201
Spline(vit) 1.000000 0.000001597 10.593968 43
Spline(mg) 0.999960 1.673807 22.737182 36

Smoothing Model Analysis
Analysis of Deviance
SourceDFSum of SquaresChi-SquarePr > ChiSq
Spline(cal_a) 3.38136 16.148270 16.1483 0.0016
Spline(cal_B) 0.14063 0.107249 0.1072 .
Spline(vit) 0.00000160 0.090490 0.0905 .
Spline(mg) 1.67381 7.585028 7.5850 0.0156

	[[alternative HTML version deleted]]