Why do we say that baseline hazard represents the hazard function of individuals with zero covariate vector always, however it will be true only for a particular case, when the independent function of covariates is exponential function. We might have some other form, in which case we will obtain the baseline hazard only when this function is unity.
Yes - that particular comment about the zero covariate vector relates specifically to the Cox model, in which the function of the covariates is an exponential function. In other cases, where we don't have an exponential function, we might still aim for a zero covariate vector yielding a value of 1 from the covariate function, for neatness' sake.
In a Cox model, you can automatically say that the baseline hazard relates to individuals with covariate values of 0. For other models, you'd need to inspect the form of the functions provided before commenting.
