手机预览点这里
2024-04-05 更新 🔗
- 写likelihood第一步，想一下 what is Y. in survival analysis， the observed outcome, is observed T + status. where T is either censoring time(C_i) or event time (T~_i)
- 两个假设很重要： 1. independent censoring (C_i and T~_i independent); 2. non-informative censoring C_i does not involve theta (params of interest)
- 因为我们观测到的结局包括了两种情况有人censor有人developed outcome，i)censored+censoring time or ii).event+event time. 那么我们就分情况写likelihood。这是因为我们的数据产生机制就是有censoring dist and event time dist (参考我们practicum是如何产生observed time的，我们是假设censoring time ~ some dist, e.g., uniform, exp； event time ～ some dist T = min{C, T~})