一致性指標（Concordance index ,又稱為C-index）經常被用來衡量模型的預測準確性，此指標能運用在未設限與設限資料下。在過去文獻中，可以使用參數模型或無母數方法來獲得區間設限資料下的一致性指標。此研究希望將回歸模型引入一致性指標中，使得模型的預測準確度更加準確，因此利用Cox模型的資訊來得到一致性指標。而當比例風險假設不成立時，在這種情況下則可以使用AFT加速失效模型作為替代。但區間設限中半母數模型的估計是困難的 ，在Cox模型下，本文引用Anderson-Bergman (2015)所提出的演算法，來估計區間設限中半母數模型參數。此外當使用AFT模型時，Gao, Zeng和Lin (2017)提出的方法。最後在模擬研究來觀察一致性指標的表現，並將該方法應用於HIV資料。;The Concordance index(C-index) was often used to measure the prediction accuracy of the model. This index can be used under uncensored and censored data. In the literature, we can use the parametric model or nonparametric method to obtain the C-index under the interval censored data. In this study, we would like to introduce the hazard regression models into the C-index to make the model prediction accuracy more efficient. The Cox model is used to incorporate into C-index due to its flexibility. Moreover, in case that proportional assumption fails, the accelerated failure time (AFT) model was used to replace the Cox model. The estimation of the semi-parametric model in interval censored data is not straighforward. Under the Cox model, we use the algorithm proposed by Anderson-Bergman (2015) to estimate the regression parameters in interval censored data. In addition, we use the method proposed by Gao, F., Zeng, D.,and Lin, D. Y. (2017) when AFT model is used. The performance of C-index is accessed by simulation study and the proposed method is applied to HIV data.
