DA_FDN2M1M2 {SPPcomb}

Data Analysis for Combining (N2,M1) + (N2,M2)

Description

The main function to solve the estimating equations constructed by combining pair (N2,M1) and (N2,M2). Since there is just one case data, no selection bias needed.

Usage

DA_FDN2M1M2(realdata_covariates, realdata_alpha, subset_2, subset_4, p, beta0)

Arguments

realdata_covariates

a list contains the following data matrics: CASEZ_2, CASEZhat_2, CASEZhat_22, CONTZ_1, CONTZhat_1, CONTZhat_2, CONTZhat_22

realdata_alpha

a list contains the following data matrics: prob_case_22, prob_cont_1, prob_cont_2, pwt_cont_2

subset_2

A vector of 1:(p-2), which is the subset of \hat{Z}_{21}, i.e. \hat{Z}_{21}^\star in equation (10) of Huang(2014). \hat{Z}_l may be highly correlated with Z_d, so it is removed in the estimation. For the view of including more information, you can use the whole dataset.

subset_4

A vector of 1:(p-2), which is the subset of \hat{Z}_{22}.

p

number of parameters.

beta0

an initial parameter for solver "nleqslv".

Details

The function solves estimating equation based on (N2,M1) and (N2,M2), see Huang(2014).

Value

A list of estimator and its standard deviation.

References

Huang, H., Ma, X., Waagepetersen, R., Holford, T.R. , Wang, R., Risch, H., Mueller, L. & Guan, Y. (2014). A New Estimation Approach for Combining Epidemiological Data From Multiple Sources, Journal of the American Statistical Association, 109:505, 11-23.

Examples

 #p <- 8
 #subset_2 <- 1:p
 #subset_4 <- 1:p
 #beta0=c(-5.4163,0.7790,-0.1289,0.2773,-0.5510,0.1568,0.4353,-0.6895)
 #DA_FDN2M1M2(realdata_covariates,realdata_alpha,subset_2,subset_4,p=p,beta0=beta0)