Exact Inference in the Proportional Hazard Model: Possibilities and Limitations

摘要: It is suggested that inference under the proportional hazard model can be carried out by programs for exact logistic regression model. Advantages of such software available and multivariate models addressed. The method has been evaluated means coverage power calculations in certain situations. In all situations was above nominal level, but on other hand rather conservative. A different type developed Type II censoring. Inference then less conservative, however there are limitations with respect to censoring mechanism, generalizations not available. This also requires extensive computational power. Performance large sample Wald, score likelihood considered. Large methods works remarkably well small data sets, statistics seems best choice. There some problems ratio may originate from how this infinite estimates parameter. Wald quite conservative very sets.