中国科学院大学学报 ›› 2026, Vol. 43 ›› Issue (3): 296-305.DOI: 10.7523/j.ucas.2024.034
收稿日期:2024-01-22
接受日期:2024-04-25
发布日期:2024-05-29
通讯作者:
张伟平
基金资助:Received:2024-01-22
Accepted:2024-04-25
Published:2024-05-29
Contact:
Weiping ZHANG
摘要:
基于Tobit回归模型,使用成对融合惩罚的正则化方法,对具有异质性的左删失数据进行子群分析,实现了回归参数估计与子群识别的同步进行。通过引入一组新变量,将原优化问题转化为可以用交替方向乘子法求解的仅含等式约束的多变量优化问题。并且,将每一步迭代的目标函数中与损失相关的多变量函数,利用广义坐标下降算法转化为一组二次优化单变量函数。证明所提算法的收敛性,并建立所得参数估计量的大样本性质。模拟研究和实际数据分析表明所提方法具有良好性能。
中图分类号:
庞珊, 张伟平. 基于成对融合惩罚的左删失数据子群分析[J]. 中国科学院大学学报, 2026, 43(3): 296-305.
Shan PANG, Weiping ZHANG. Subgroup analysis for left-censored data based on pairwise fusion penalty[J]. Journal of University of Chinese Academy of Sciences, 2026, 43(3): 296-305.
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.033 4(0.022 0) | 0.028 8(0.018 6) | 0.028 0(0.008 6) | 0.027 6(0.009 7) | 0.039 8(0.023 9) |
| MCP | 0.033 4(0.022 3) | 0.027 9(0.018 1) | 0.028 3(0.008 8) | 0.027 6(0.009 7) | 0.039 0(0.021 4) | |
| L1 | 1.224 8(0.000 5) | 1.003 4(0.002 4) | 0.104 1(0.037 5) | 0.093 2(0.028 0) | 0.475 6(0.066 9) | |
| Oracle | 0.018 9(0.027 4) | 0.019 9(0.031 7) | 0.022 9(0.013 2) | 0.018 6(0.008 1) | 0.019 4(0.013 5) | |
| 1.5 | SCAD | 0.034 1(0.022 3) | 0.029 4(0.018 9) | 0.027 6(0.008 6) | 0.023 6(0.007 5) | 0.039 4(0.020 5) |
| MCP | 0.032 4(0.021 4) | 0.030 4(0.020 2) | 0.027 9(0.008 7) | 0.022 9(0.007 3) | 0.040 2(0.020 3) | |
| L1 | 0.816 5(0.000 5) | 0.668 9(0.001 6) | 0.087 8(0.032 6) | 0.068 2(0.020 2) | 0.185 1(0.054 4) | |
| Oracle | 0.019 1(0.029 4) | 0.019 8(0.023 8) | 0.020 8(0.013 2) | 0.016 0(0.008 2) | 0.019 1(0.013 8) | |
| 2 | SCAD | 0.032 6(0.020 8) | 0.028 0(0.017 7) | 0.027 7(0.008 9) | 0.023 7(0.006 8) | 0.034 1(0.018 2) |
| MCP | 0.033 5(0.021 4) | 0.028 6(0.018 0) | 0.027 8(0.008 7) | 0.023 0(0.007 5) | 0.033 6(0.017 7) | |
| L1 | 0.668 0(0.000 9) | 0.641 1(0.000 6) | 0.082 9(0.018 5) | 0.082 2(0.018 3) | 0.150 3(0.051 0) | |
| Oracle | 0.019 8(0.029 5) | 0.017 7(0.030 6) | 0.016 5(0.038 6) | 0.012 5(0.030 3) | 0.018 8(0.013 4) | |
| 1 | SCAD | 0.036 8(0.023 4) | 0.031 6(0.018 0) | 0.028 9(0.010 5) | 0.026 1(0.008 5) | 0.043 6(0.025 5) |
| MCP | 0.040 3(0.023 3) | 0.035 1(0.020 2) | 0.034 7(0.011 8) | 0.021 9(0.007 5) | 0.043 6(0.025 5) | |
| L1 | 1.633 1(0.000 4) | 1.338 7(0.003 7) | 0.132 4(0.039 3) | 0.110 1(0.036 0) | 0.935 8(0.058 6) | |
| Oracle | 0.020 7(0.029 0) | 0.020 3(0.029 4) | 0.025 4(0.010 0) | 0.020 7(0.011 3) | 0.018 2(0.010 8) | |
| 1.5 | SCAD | 0.036 3(0.020 9) | 0.031 6(0.019 9) | 0.034 5(0.011 4) | 0.028 7(0.010 1) | 0.043 5(0.022 0) |
| MCP | 0.040 1(0.026 2) | 0.034 6(0.022 7) | 0.034 5(0.011 4) | 0.028 7(0.010 1) | 0.044 0(0.021 9) | |
| L1 | 1.474 9(0.000 1) | 1.336 1(0.002 0) | 0.127 3(0.046 9) | 0.106 3(0.039 9) | 0.792 2(0.064 6) | |
| Oracle | 0.021 6(0.024 3) | 0.020 3(0.028 3) | 0.024 5(0.011 1) | 0.021 3(0.011 2) | 0.017 4(0.011 1) | |
| 2 | SCAD | 0.038 2(0.023 0) | 0.033 0(0.019 6) | 0.033 6(0.011 6) | 0.027 5(0.009 7) | 0.034 1(0.021 1) |
| MCP | 0.031 2(0.020 0) | 0.031 0(0.021 4) | 0.033 2(0.012 1) | 0.027 2(0.008 0) | 0.034 1(0.021 1) | |
| L1 | 1.224 7(0.000 2) | 1.001 9(0.001 3) | 0.071 6(0.025 4) | 0.059 8(0.021 9) | 0.372 3(0.058 7) | |
| Oracle | 0.019 4(0.025 2) | 0.018 2(0.029 4) | 0.023 8(0.010 1) | 0.020 2(0.010 5) | 0.017 0(0.010 9) | |
表1 参数估计量RMSE和MAE的平均值和标准差
Table 1 Mean and standard deviation of RMSE and MAE of parameter estimators
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.033 4(0.022 0) | 0.028 8(0.018 6) | 0.028 0(0.008 6) | 0.027 6(0.009 7) | 0.039 8(0.023 9) |
| MCP | 0.033 4(0.022 3) | 0.027 9(0.018 1) | 0.028 3(0.008 8) | 0.027 6(0.009 7) | 0.039 0(0.021 4) | |
| L1 | 1.224 8(0.000 5) | 1.003 4(0.002 4) | 0.104 1(0.037 5) | 0.093 2(0.028 0) | 0.475 6(0.066 9) | |
| Oracle | 0.018 9(0.027 4) | 0.019 9(0.031 7) | 0.022 9(0.013 2) | 0.018 6(0.008 1) | 0.019 4(0.013 5) | |
| 1.5 | SCAD | 0.034 1(0.022 3) | 0.029 4(0.018 9) | 0.027 6(0.008 6) | 0.023 6(0.007 5) | 0.039 4(0.020 5) |
| MCP | 0.032 4(0.021 4) | 0.030 4(0.020 2) | 0.027 9(0.008 7) | 0.022 9(0.007 3) | 0.040 2(0.020 3) | |
| L1 | 0.816 5(0.000 5) | 0.668 9(0.001 6) | 0.087 8(0.032 6) | 0.068 2(0.020 2) | 0.185 1(0.054 4) | |
| Oracle | 0.019 1(0.029 4) | 0.019 8(0.023 8) | 0.020 8(0.013 2) | 0.016 0(0.008 2) | 0.019 1(0.013 8) | |
| 2 | SCAD | 0.032 6(0.020 8) | 0.028 0(0.017 7) | 0.027 7(0.008 9) | 0.023 7(0.006 8) | 0.034 1(0.018 2) |
| MCP | 0.033 5(0.021 4) | 0.028 6(0.018 0) | 0.027 8(0.008 7) | 0.023 0(0.007 5) | 0.033 6(0.017 7) | |
| L1 | 0.668 0(0.000 9) | 0.641 1(0.000 6) | 0.082 9(0.018 5) | 0.082 2(0.018 3) | 0.150 3(0.051 0) | |
| Oracle | 0.019 8(0.029 5) | 0.017 7(0.030 6) | 0.016 5(0.038 6) | 0.012 5(0.030 3) | 0.018 8(0.013 4) | |
| 1 | SCAD | 0.036 8(0.023 4) | 0.031 6(0.018 0) | 0.028 9(0.010 5) | 0.026 1(0.008 5) | 0.043 6(0.025 5) |
| MCP | 0.040 3(0.023 3) | 0.035 1(0.020 2) | 0.034 7(0.011 8) | 0.021 9(0.007 5) | 0.043 6(0.025 5) | |
| L1 | 1.633 1(0.000 4) | 1.338 7(0.003 7) | 0.132 4(0.039 3) | 0.110 1(0.036 0) | 0.935 8(0.058 6) | |
| Oracle | 0.020 7(0.029 0) | 0.020 3(0.029 4) | 0.025 4(0.010 0) | 0.020 7(0.011 3) | 0.018 2(0.010 8) | |
| 1.5 | SCAD | 0.036 3(0.020 9) | 0.031 6(0.019 9) | 0.034 5(0.011 4) | 0.028 7(0.010 1) | 0.043 5(0.022 0) |
| MCP | 0.040 1(0.026 2) | 0.034 6(0.022 7) | 0.034 5(0.011 4) | 0.028 7(0.010 1) | 0.044 0(0.021 9) | |
| L1 | 1.474 9(0.000 1) | 1.336 1(0.002 0) | 0.127 3(0.046 9) | 0.106 3(0.039 9) | 0.792 2(0.064 6) | |
| Oracle | 0.021 6(0.024 3) | 0.020 3(0.028 3) | 0.024 5(0.011 1) | 0.021 3(0.011 2) | 0.017 4(0.011 1) | |
| 2 | SCAD | 0.038 2(0.023 0) | 0.033 0(0.019 6) | 0.033 6(0.011 6) | 0.027 5(0.009 7) | 0.034 1(0.021 1) |
| MCP | 0.031 2(0.020 0) | 0.031 0(0.021 4) | 0.033 2(0.012 1) | 0.027 2(0.008 0) | 0.034 1(0.021 1) | |
| L1 | 1.224 7(0.000 2) | 1.001 9(0.001 3) | 0.071 6(0.025 4) | 0.059 8(0.021 9) | 0.372 3(0.058 7) | |
| Oracle | 0.019 4(0.025 2) | 0.018 2(0.029 4) | 0.023 8(0.010 1) | 0.020 2(0.010 5) | 0.017 0(0.010 9) | |
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 3.03(0.222 7) | 0.998 5(0.010 5) | 0.996 5(0.025 3) | 98 |
| MCP | 3.03(0.258 1) | 0.998 4(0.011 7) | 0.996 3(0.028 4) | 98 | |
| L1 | 1(0) | 0.332 2(0) | 0(0) | 0 | |
| 1.5 | SCAD | 3.02(0.141 4) | 0.998 9(0.007 2) | 0.997 5(0.017 0) | 98.6 |
| MCP | 3.03 (0.222 7) | 0.998 5(0.010 5) | 0.996 5(0.025 3) | 98.2 | |
| L1 | 1(0) | 0.332 2 (0) | 0(0) | 0 | |
| 2 | SCAD | 3.02(0.2) | 0.999 0(0.009 0) | 0.997 7(0.022 0) | 99 |
| MCP | 3.01(0.1) | 0.999 4(0.005 1) | 0.998 7(0.012 0) | 99 | |
| L1 | 1(0) | 0.332 2(0) | 0(0) | 0 | |
| 1 | SCAD | 3.18(0.542 9) | 0.990 7(0.028 7) | 0.977 4(0.070 5) | 97 |
| MCP | 3.2(0.556 6) | 0.990 2(0.029 5) | 0.976 1(0.072 3) | 96.6 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
| 1.5 | SCAD | 3.17(0.551 4) | 0.991 0(0.029 4) | 0.978 1(0.073 1) | 98.2 |
| MCP | 3.18(0.557 4) | 0.990 5(0.029 7) | 0.976 8(0.073 8) | 98 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
| 2 | SCAD | 3.04(0.242 8) | 0.998 0(0.011 8) | 0.995 2(0.028 2) | 98.8 |
| MCP | 3.05(0.286 7) | 0.997 5(0.013 5) | 0.994 2(0.032 5) | 98.6 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
表2 分组结果
Table 2 Grouping results
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 3.03(0.222 7) | 0.998 5(0.010 5) | 0.996 5(0.025 3) | 98 |
| MCP | 3.03(0.258 1) | 0.998 4(0.011 7) | 0.996 3(0.028 4) | 98 | |
| L1 | 1(0) | 0.332 2(0) | 0(0) | 0 | |
| 1.5 | SCAD | 3.02(0.141 4) | 0.998 9(0.007 2) | 0.997 5(0.017 0) | 98.6 |
| MCP | 3.03 (0.222 7) | 0.998 5(0.010 5) | 0.996 5(0.025 3) | 98.2 | |
| L1 | 1(0) | 0.332 2 (0) | 0(0) | 0 | |
| 2 | SCAD | 3.02(0.2) | 0.999 0(0.009 0) | 0.997 7(0.022 0) | 99 |
| MCP | 3.01(0.1) | 0.999 4(0.005 1) | 0.998 7(0.012 0) | 99 | |
| L1 | 1(0) | 0.332 2(0) | 0(0) | 0 | |
| 1 | SCAD | 3.18(0.542 9) | 0.990 7(0.028 7) | 0.977 4(0.070 5) | 97 |
| MCP | 3.2(0.556 6) | 0.990 2(0.029 5) | 0.976 1(0.072 3) | 96.6 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
| 1.5 | SCAD | 3.17(0.551 4) | 0.991 0(0.029 4) | 0.978 1(0.073 1) | 98.2 |
| MCP | 3.18(0.557 4) | 0.990 5(0.029 7) | 0.976 8(0.073 8) | 98 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
| 2 | SCAD | 3.04(0.242 8) | 0.998 0(0.011 8) | 0.995 2(0.028 2) | 98.8 |
| MCP | 3.05(0.286 7) | 0.997 5(0.013 5) | 0.994 2(0.032 5) | 98.6 | |
| L1 | 1(0) | 0.331 1(0) | 0(0) | 0 | |
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.040 1(0.023 2) | 0.034 4(0.019 6) | 0.034 7(0.012 7) | 0.029 3(0.010 9) | 0.032 6(0.019 7) |
| MCP | 0.041 0(0.022 9) | 0.035 2(0.019 3) | 0.035 0(0.012 1) | 0.029 5(0.010 7) | 0.032 4(0.019 6) | |
| Oracle | 0.029 4(0.025 2) | 0.021 5(0.023 0) | 0.018 0(0.013 9) | 0.011 6(0.011 8) | 0.021 4(0.014 2) | |
| 1.5 | SCAD | 0.041 1(0.026 2) | 0.035 0(0.022 2) | 0.034 8(0.012 3) | 0.029 3(0.010 5) | 0.039 0(0.022 0) |
| MCP | 0.041 3(0.026 3) | 0.035 2(0.022 2) | 0.034 8(0.012 3) | 0.029 3(0.010 5) | 0.039 0(0.021 9) | |
| Oracle | 0.028 9(0.024 3) | 0.021 1(0.022 2) | 0.017 9(0.013 1) | 0.011 7(0.011 1) | 0.021 1(0.014 3) | |
表3 混合正态误差下参数估计量的RMSE和MAE平均值和标准差
Table 3 Mean and standard deviation of RMSE and MAE of parameter estimators with mixed normal error
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.040 1(0.023 2) | 0.034 4(0.019 6) | 0.034 7(0.012 7) | 0.029 3(0.010 9) | 0.032 6(0.019 7) |
| MCP | 0.041 0(0.022 9) | 0.035 2(0.019 3) | 0.035 0(0.012 1) | 0.029 5(0.010 7) | 0.032 4(0.019 6) | |
| Oracle | 0.029 4(0.025 2) | 0.021 5(0.023 0) | 0.018 0(0.013 9) | 0.011 6(0.011 8) | 0.021 4(0.014 2) | |
| 1.5 | SCAD | 0.041 1(0.026 2) | 0.035 0(0.022 2) | 0.034 8(0.012 3) | 0.029 3(0.010 5) | 0.039 0(0.022 0) |
| MCP | 0.041 3(0.026 3) | 0.035 2(0.022 2) | 0.034 8(0.012 3) | 0.029 3(0.010 5) | 0.039 0(0.021 9) | |
| Oracle | 0.028 9(0.024 3) | 0.021 1(0.022 2) | 0.017 9(0.013 1) | 0.011 7(0.011 1) | 0.021 1(0.014 3) | |
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 3(0) | 1(0) | 1(0) | 100 |
| MCP | 3(0) | 1(0) | 1(0) | 100 | |
| 1.5 | SCAD | 3(0) | 1(0) | 1(0) | 100 |
| MCP | 3(0) | 1(0) | 1(0) | 100 |
表4 混合正态误差下的分组结果
Table 4 Grouping results with mixed normal error
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 3(0) | 1(0) | 1(0) | 100 |
| MCP | 3(0) | 1(0) | 1(0) | 100 | |
| 1.5 | SCAD | 3(0) | 1(0) | 1(0) | 100 |
| MCP | 3(0) | 1(0) | 1(0) | 100 |
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.037 3(0.020 4) | 0.032 0(0.017 6) | 0.028 6(0.009 9) | 0.023 9(0.008 7) | 0.041 9(0.023 8) |
| MCP | 0.037 4(0.020 2) | 0.032 0(0.017 5) | 0.028 8(0.009 8) | 0.024 0(0.008 4) | 0.042 1(0.023 8) | |
| Oracle | 0.027 3(0.018 5) | 0.018 2(0.016 1) | 0.010 8(0.010 7) | 0.015 8(0.009 2) | 0.016 8(0.011 8) | |
| 1.5 | SCAD | 0.033 6(0.018 2) | 0.028 7(0.015 8) | 0.028 1(0.009 7) | 0.023 2(0.008 5) | 0.046 7(0.026 1) |
| MCP | 0.033 6(0.017 6) | 0.028 7(0.015 4) | 0.028 2(0.009 6) | 0.023 4(0.008 1) | 0.046 9(0.026 0) | |
| Oracle | 0.022 3(0.019 1) | 0.017 8(0.016 6) | 0.019 8(0.010 4) | 0.014 7(0.009 1) | 0.015 9(0.011 7) | |
表5 K=5时,参数估计量的RMSE和MAE平均值和标准差
Table 5 Mean and standard deviation of RMSE and MAE of parameter estimators when K=5
| RMSE | MAE | RMSE | MAE | MAE | ||
|---|---|---|---|---|---|---|
| 1 | SCAD | 0.037 3(0.020 4) | 0.032 0(0.017 6) | 0.028 6(0.009 9) | 0.023 9(0.008 7) | 0.041 9(0.023 8) |
| MCP | 0.037 4(0.020 2) | 0.032 0(0.017 5) | 0.028 8(0.009 8) | 0.024 0(0.008 4) | 0.042 1(0.023 8) | |
| Oracle | 0.027 3(0.018 5) | 0.018 2(0.016 1) | 0.010 8(0.010 7) | 0.015 8(0.009 2) | 0.016 8(0.011 8) | |
| 1.5 | SCAD | 0.033 6(0.018 2) | 0.028 7(0.015 8) | 0.028 1(0.009 7) | 0.023 2(0.008 5) | 0.046 7(0.026 1) |
| MCP | 0.033 6(0.017 6) | 0.028 7(0.015 4) | 0.028 2(0.009 6) | 0.023 4(0.008 1) | 0.046 9(0.026 0) | |
| Oracle | 0.022 3(0.019 1) | 0.017 8(0.016 6) | 0.019 8(0.010 4) | 0.014 7(0.009 1) | 0.015 9(0.011 7) | |
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 5(0) | 1(0) | 1(0) | 100 |
| MCP | 5(0) | 1(0) | 1(0) | 100 | |
| 1.5 | SCAD | 5(0) | 1(0) | 1(0) | 100 |
| MCP | 5(0) | 1(0) | 1(0) | 100 |
表6 K=5时的分组结果
Table 6 Grouping results when K=5
| RI | ARI | ||||
|---|---|---|---|---|---|
| 1 | SCAD | 5(0) | 1(0) | 1(0) | 100 |
| MCP | 5(0) | 1(0) | 1(0) | 100 | |
| 1.5 | SCAD | 5(0) | 1(0) | 1(0) | 100 |
| MCP | 5(0) | 1(0) | 1(0) | 100 |
| RI | ARI | ||
|---|---|---|---|
| Tobit SCAD | 2 | 0.924 | 0.771 |
| Tobit MCP | 2 | 0.924 | 0.771 |
表7 分组准确度
Table 7 Grouping accuracy
| RI | ARI | ||
|---|---|---|---|
| Tobit SCAD | 2 | 0.924 | 0.771 |
| Tobit MCP | 2 | 0.924 | 0.771 |
| Tobit SCAD | 2.720 | 5.073 | -0.098 | -0.138 | -0.092 | 0.134 |
| Tobit MCP | 2.744 | 5.090 | -0.102 | -0.146 | -0.093 | 0.136 |
| Oracle | 2.694 | 5.450 | -0.072 | -0.133 | -0.089 | 0.158 |
表8 不同方法的参数估计值
Table 8 Parameter estimates for different methods
| Tobit SCAD | 2.720 | 5.073 | -0.098 | -0.138 | -0.092 | 0.134 |
| Tobit MCP | 2.744 | 5.090 | -0.102 | -0.146 | -0.093 | 0.136 |
| Oracle | 2.694 | 5.450 | -0.072 | -0.133 | -0.089 | 0.158 |
| Tobit损失 | MSE | |
|---|---|---|
| Tobit SCAD | 1.290(0.038) | 0.053(0.018) |
| Tobit MCP | 1.290(0.039) | 0.068(0.023) |
| Oracle | 0.874(0.027) | 0.044(0.014) |
表9 不同方法的Tobit损失和预测MSE
Table 9 Tobit losses and predicted MSEs for different methods
| Tobit损失 | MSE | |
|---|---|---|
| Tobit SCAD | 1.290(0.038) | 0.053(0.018) |
| Tobit MCP | 1.290(0.039) | 0.068(0.023) |
| Oracle | 0.874(0.027) | 0.044(0.014) |
| [1] | Tobin J. Estimation of relationships for limited dependent variables[J]. Econometrica, 1958, 26(1): 24. DOI: 10.2307/1907382 . |
| [2] | Buckley J, James I. Linear regression with censored data[J]. Biometrika, 1979, 66(3): 429-436. DOI: 10.1093/biomet/66.3.429 . |
| [3] | Miller R G. Least squares regression with censored data[J]. Biometrika, 1976, 63(3): 449-464. DOI: 10.1093/biomet/63.3.449 . |
| [4] | Wei L J, Ying Z, Lin D Y. Linear regression analysis of censored survival data based on rank tests[J]. Biometrika, 1990, 77(4): 845-851. DOI: 10.1093/biomet/77.4.845 . |
| [5] | Prentice R L. Linear rank tests with right censored data[J]. Biometrika, 1978, 65(1): 167-179. DOI: 10.1093/biomet/65.1.167 . |
| [6] | 陈菲菲, 孙志华, 叶雪. 失效信息随机缺失时可加危险率模型的统计推断[J]. 中国科学院大学学报, 2016, 33(4): 443-453. DOI: 10.7523/j.issn.2095-6134.2016.04.003 . |
| [7] | Cuzick J. Asymptotic properties of censored linear rank tests[J]. The Annals of Statistics, 1985, 13(1): 133-141. DOI: 10.1214/aos/1176346581 . |
| [8] | Fay M P, Shaw P A. Exact and asymptotic weighted logrank tests for interval censored data: the interval r package[J]. Journal of Statistical Software, 2010, 36(2): 1-34. DOI: 10.18637/jss.v036.i02 . |
| [9] | Amemiya T. Regression analysis when the dependent variable is truncated normal[J]. Econometrica, 1973, 41(6): 997. DOI: 10.2307/1914031 . |
| [10] | Olsen R J. Note on the uniqueness of the maximum likelihood estimator for the Tobit model[J]. Econometrica, 1978, 46(5): 1211. DOI: 10.2307/1911445 . |
| [11] | Amemiya T. Tobit models: a survey[J]. Journal of Econometrics, 1984, 24(1/2): 3-61. DOI: 10.1016/0304-4076(84)90074-5 . |
| [12] | Turnbull B W. The empirical distribution function with arbitrarily grouped, censored and truncated data[J]. Journal of the Royal Statistical Society Series B: Statistical Methodology, 1976, 38(3): 290-295. DOI: 10.1111/j.2517-6161.1976.tb01597.x . |
| [13] | Zhou M. Empirical likelihood ratio with arbitrarily censored/truncated data by EM algorithm[J]. Journal of Computational and Graphical Statistics, 2005, 14(3): 643-656. DOI: 10.1198/106186005x59270 . |
| [14] | Everitt B S, Hand D J. Finite mixture distributions[M]. Dordrecht: Springer Netherlands, 1981. DOI: 10.1007/978-94-009-5897-5 . |
| [15] | Wei S S, Kosorok M R. Latent supervised learning[J]. Journal of the American Statistical Association, 2013, 108(503): 957-970. DOI: 10.1080/01621459.2013.789695 . |
| [16] | Shen J, He X M. Inference for subgroup analysis with a structured logistic-normal mixture model[J]. Journal of the American Statistical Association, 2015, 110(509): 303-312. DOI: 10.1080/01621459.2014.894763 . |
| [17] | Wu R F, Zheng M, Yu W. Subgroup analysis with time-to-event data under a logistic-cox mixture model[J]. Scandinavian Journal of Statistics, 2016, 43(3): 863-878. DOI: 10.1111/sjos.12213 . |
| [18] | Ma S J, Huang J. A concave pairwise fusion approach to subgroup analysis[J]. Journal of the American Statistical Association, 2017, 112(517): 410-423. DOI: 10.1080/01621459.2016.1148039 . |
| [19] | Ma S J, Huang J, Zhang Z W, et al. Exploration of heterogeneous treatment effects via concave fusion[J]. The International Journal of Biostatistics, 2020, 16(1): arXiv: 1607.03717. DOI: 10.1515/ijb-2018-0026 . |
| [20] | Chen J X, Tran-Dinh Q, Kosorok M R, et al. Identifying heterogeneous effect using latent supervised clustering with adaptive fusion[J]. Journal of Computational and Graphical Statistics: a Joint Publication of American Statistical Association, Institute of Mathematical Statistics, Interface Foundation of North America, 2021, 30(1): 43-54. DOI: 10.1080/10618600.2020.1763808 . |
| [21] | 许赵辉, 郑泽敏, 吴捷. 基于加速失效模型的子群分析方法[J]. 应用概率统计, 2023, 39(5): 765-780. DOI: 10.3969/j.issn.1001-4268.2023.05.010 . |
| [22] | Fan J Q, Li R Z. Variable selection via nonconcave penalized likelihood and its oracle properties[J]. Journal of the American Statistical Association, 2001, 96(456): 1348-1360. DOI: 10.1198/016214501753382273 . |
| [23] | Zhang C H. Nearly unbiased variable selection under minimax concave penalty[J]. The Annals of Statistics, 2010, 38(2): 894-942. DOI: 10.1214/09-aos729 . |
| [24] | Yang Y, Zou H. An efficient algorithm for computing the HHSVM and its generalizations[J]. Journal of Computational and Graphical Statistics, 2013, 22(2): 396-415. DOI: 10.1080/10618600.2012.680324 . |
| [25] | Wang H S, Li R Z, Tsai C L. Tuning parameter selectors for the smoothly clipped absolute deviation method[J]. Biometrika, 2007, 94(3): 553-568. DOI: 10.1093/biomet/asm053 . |
| [26] | Wang H S, Li B, Leng C L. Shrinkage tuning parameter selection with a diverging number of parameters[J]. Journal of the Royal Statistical Society Series B: Statistical Methodology, 2009, 71(3): 671-683. DOI: 10.1111/j.1467-9868.2008.00693.x . |
| [27] | Yan X D, Yin G S, Zhao X Q. Subgroup analysis in censored linear regression[J]. Statistica Sinica, 2021: 1027-1054. DOI: 10.5705/ss.202018.0319 . |
| [28] | Miller R, Halpern J. Regression with censored data[J]. Biometrika, 1982, 69(3): 521-531. DOI: 10.1093/biomet/69.3.521 . |
| [29] | Gandhi R T, Tashima K T, Smeaton L M, et al. Long-term outcomes in a large randomized trial of HIV-1 salvage therapy: 96-week results of AIDS clinical trials group a5241 (options)[J]. The Journal of Infectious Diseases, 2020, 221(9): 1407-1415. DOI: 10.1093/infdis/jiz281 . |
| [30] | Jacobson T, Zou H. High-dimensional censored regression via the penalized tobit likelihood[J]. Journal of Business & Economic Statistics, 2024, 42(1): 286-297. DOI: 10.1080/07350015.2023.2182309 . |
| [1] | 吴慧桢, 张三国. 基于模型平均与γ-散度的稳健半监督学习方法[J]. 中国科学院大学学报, 2026, 43(1): 14-22. |
| [2] | 翟浩然, 张三国. 引入集群效应的跨领域推荐新方法[J]. 中国科学院大学学报, 2025, 42(2): 153-158. |
| [3] | 张晓灵, 任明旸, 张三国. 稳健的个体化亚组分析[J]. 中国科学院大学学报, 2024, 41(2): 151-164. |
| [4] | 张家睿, 吴耀华. 高维生存分析数据在带有测量误差情形下的变量选择方法[J]. 中国科学院大学学报, 2023, 40(1): 12-20. |
| [5] | 刘琰, 李仕明, 张三国. 基于符号秩的高维均值检验[J]. 中国科学院大学学报, 2022, 39(5): 586-592. |
| [6] | 董玉林, 郭潇. 惩罚逻辑回归的击穿点[J]. 中国科学院大学学报, 2020, 37(5): 582-592. |
| [7] | 陈菲菲, 孙志华, 叶雪. 失效信息随机缺失时可加危险率模型的统计推断[J]. 中国科学院大学学报, 2016, 33(4): 443-453. |
| [8] | 钞婷, 李启寨, 刘卓军, 孙才, 孙云刚. 基于广义和校准马氏距离对IP地址威胁程度的诊断[J]. 中国科学院大学学报, 2015, 32(1): 18-24. |
| [9] | 黄传明, 张晓旭, 张三国. 基于信息量准则的广义Lambda分布变点分析[J]. 中国科学院大学学报, 2013, 30(6): 728-736. |
| [10] | 仇丽莎, 韦来生. 正态总体均值和误差方差同时的经验Bayes估计[J]. 中国科学院大学学报, 2013, 30(4): 454-461. |
| [11] | 李翔, 韦来生. 指数分布定数截尾数据下刻度参数的经验Bayes估计[J]. 中国科学院大学学报, 2011, 28(2): 147-154. |
| [12] | 王鹏远, 韦来生. 回归系数一类线性估计的小样本性质[J]. 中国科学院大学学报, 2009, 26(3): 296-302. |
| [13] | 扈慧敏 杨荣 徐兴忠. 单因素方差分析模型中的广义p-值[J]. 中国科学院大学学报, 2007, 24(4): 408-418. |
| [14] | 李海明. 中国水稻品种改良以及对水稻生产的影响[J]. 中国科学院大学学报, 2007, 24(1): 1-8. |
| [15] | 魏 莉 孔胜春 韦来生. 刻度指数族参数的经验Bayes检验的收敛速度[J]. 中国科学院大学学报, 2007, 24(1): 9-17. |
| 阅读次数 | ||||||
|
全文 |
|
|||||
|
摘要 |
|
|||||
