Data and code for "Exact Finite-Sample Inference for the Exponential Proportional Hazards Model via the Empirical Characteristic Function"

Reproducibility package for the article submitted to Computational Statistics and Data Analysis. Contains six Python scripts that reproduce every table and closed-form result, the leukemia remission dataset (Freireich et al., 1963) used in the real-data example, and pre-computed outputs. Scripts cover Monte Carlo validation of the exact finite-sample bias/variance formulas, coverage of the exact confidence intervals, the Weibull PHM extension (known and unknown shape), a Cauchy ECF-vs-ML comparison, misspecification and dependent-censoring analyses, and a symbolic verification of all closed-form formulas. All results are fully reproducible with numpy, scipy and sympy under fixed random seeds.本数据集为投稿至《Computational Statistics and Data Analysis》的学术论文配套可复现性研究包。内含6个Python脚本，可复现全部表格与闭式解析结果；附带实证分析环节所用的白血病缓解数据集（Leukemia Remission Dataset，Freireich等，1963），以及预计算输出结果。脚本涵盖以下研究内容：精确有限样本偏差与方差公式的蒙特卡洛验证、精确置信区间的覆盖率检验、威布尔比例风险模型（Weibull PHM）扩展分析（含形状参数已知与未知两种场景）、柯西经验特征函数（Empirical Characteristic Function，ECF）与极大似然估计（Maximum Likelihood，ML）对比分析、模型误设定与相依删失分析，以及所有闭式解析公式的符号验证。所有结果均可在固定随机种子的前提下，借助numpy、scipy与sympy库实现完全复现。