Real-time monitored binary data are often recorded along with a large amount of associated covariates for biomedical image processing. Serially measured binary outcomes and covariates could be autocorrelated. Appropriate variable selection schemes are necessary to find a set of influential covariates on the changes in the correlated binary outcomes. Selected variables can be used as feedback information to reduce the dimension of the database. In this context, we examine the performance of the stepwise correlated binary regression. Several realistic situations of the real-time monitored binary data are considered in Monte-Carlo simulation. Results of a simulation study are discussed.
Bibliographical noteFunding Information:
This work was supported by the Yonsei University Research Fund of 1997.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- Health Informatics