how to interpret chi sq, P(chi) and calculate CI for coef of gam functions

["Poh Siew Hua" <ESHPoh@ntu.edu.sg>]

Hi list,
 
 
Could anyone help me on this? How do we interpret the chi sq and P(chi) values 
for whether to include the lo terms or not.   Which df do we use, the NPAR DF 
or just the DF? How do we calculate the confidence intervals for the 
coefficients of the gam function?
 
Thanks!
Rgds,
Poh Siew Hua
 
 
 
> mfit <- gam(Mortality.Counts ~ lo(day, 0.15) + lo(Temperature, 0.55) + lo(
 T1Lag, 0.2) + lo(humidity, 0.95) , na.action = 
 na.omit, family = poisson(link = log), data = final3.data)
> coef(mfit)
 (Intercept) lo(day, 0.15) lo(Temperature, 0.55) lo(T1Lag, 0.2) 
    3.687732    -0.1181391            -0.7309892       2.417947
 lo(humidity, 0.95) 
> summary.gam(mfit)
Call: gam(formula = Mortality.Counts ~ lo(day, 0.15) + lo(Temperature, 0.55) + 
lo(
 T1Lag, 0.2) + lo(humidity, 0.95) , family = 
 poisson(link = log), data = final3.data, na.action = na.omit)
Deviance Residuals:
       Min         1Q     Median        3Q     Max 
 -4.014915 -0.7013316 -0.0171393 0.6563786 3.35557
(Dispersion Parameter for Poisson family taken to be 1 )
    Null Deviance: 2076.159 on 1731 degrees of freedom
Residual Deviance: 1749.195 on 1672.452 degrees of freedom
2 observations deleted due to missing values 
Number of Local Scoring Iterations: 3 
DF for Terms and Chi-squares for Nonparametric Effects
                      Df Npar Df Npar Chisq    P(Chi) 
          (Intercept)  1                             
        lo(day, 0.15)  1    10.5   97.87262 0.0000000
lo(Temperature, 0.55)  1     2.0    1.40289 0.4951667
       lo(T1Lag, 0.2)  1     9.2   54.91880 0.0000000
   lo(humidity, 0.95)  1     0.5    0.24081 0.3629457