Why the parameters optimized by Bayesian parameters are not optimal？

liupei101 / TFDeepSurv

COX Proportional risk model and survival analysis implemented by tensorflow.

MIT License

99 stars 27 forks source link

Why the parameters optimized by Bayesian parameters are not optimal？ #18

Closed DWang12138 closed 1 year ago

DWang12138 commented 1 year ago

The training set is adjusted according to the proportion of 0.8, and the parameters after Bayesian parameter optimization are not optimal for the whole training set. Do you know how to solve this problem?

liupei101 commented 1 year ago

Bayesian parameter optimization cannot always find the optimal solution. It's just utilized to help us obtain some model hyper-parameters that are expected to be good (or not too bad) for fitting the desired model.