本篇考慮使用線性轉換模型分析不易感受性區間設限資料，在未知基底風險函數(Unknown baseline hazard function)的估計利用核函數(Kernel)做平滑估計，並考慮EM演算法和牛頓迭代法估計參數，並經由模擬驗證。 As time progresses, continuous development, there are more and more interval censoring data with clinical trials. Sometimes, it is hard to observe the exact time of event, but we know the observed failure time falls within a time period. In this thesis, we consider mixture cure models for interval censored data with a cured subgroup, where subjects in this subgroup are not susceptible to the event of interest. We suppose logistic regression to estimate cure proportion.
In addition, we consider semiparametric transformation models to analysis the event data. We focus on reparametrizing the step function of unknown baseline hazard function by the logarithm of its jump sizes in Chapter 3, and a kernel-based approach for smooth estimation of unknown baseline hazard function in Chapter 4. The EM algorithm is developed for the estimation and simulation studies are conducted.
