What is the assumption we must check if Cox proportional hazard regression is used？
Is my step and method to build on Cox proportional hazard model correct, or not?
- Can I used ln（-ln（survival））curve check PH assumption for categorical covariate, Schoenfeld residuals test for continuous covariate
- Maximum partial likelihood estimates is used to estimate coefficient
- After that use Wald test check significance of coefficients
- Stepwise is used to select covariate
- Overall goodness of fit of Cox proportional hazard model is check using Cox-Snell residual plot