Poisson分布下基于梯度统计量的慢性病风险比的置信区间构造

Confidence Interval Construction of Risk Ratio of Chronic Diseases Based on Gradient Statistics under Poisson Distribution

摘要: 结合Poisson抽样的优点以及慢性病发病周期长和发病率低的特点，利用梯度统计量的方法来构造Poisson分布下风险比的置信区间。该方法的优点在于不需要计算Fisher信息矩阵及其逆矩阵，与得分统计量方法相比，这大大简化了计算过程。同时，通过实例和蒙特卡洛模拟，与传统的置信区间构造方法进行比较。结果表明，基于梯度统计量的构建方法可以得到较好的覆盖概率和较短的区间长度。

Abstract: Combining the advantages of Poisson sampling and the characteristics of long occurrence period and low incidence of chronic diseases, the confidence interval of risk ratio under Poisson distribu-tion is constructed by using gradient statistics. The advantage of this method is that it does not need to calculate Fisher information matrix and its inverse matrix, which greatly simplifies the calculation process compared with the score statistics method. At the same time, an example and Monte Carlo simulation are used. Compared with the traditional confidence interval construction method, the results show that the method based on gradient statistics can obtain better coverage probability and shorter interval length.