《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (7): 2153-2161.DOI: 10.11772/j.issn.1001-9081.2024070942
• CCF第39届中国计算机应用大会 (CCF NCCA 2024) • 上一篇 下一篇
收稿日期:
2024-07-08
修回日期:
2024-09-05
接受日期:
2024-10-09
发布日期:
2025-07-10
出版日期:
2025-07-10
通讯作者:
何玉林
作者简介:
陈佳琪(1999—),女,广东普宁人,博士研究生,主要研究方向:多样本统计分析、数据挖掘、机器学习基金资助:
Jiaqi CHEN1,2, Yulin HE2(), Yingchao CHENG2, Zhexue HUANG1,2
Received:
2024-07-08
Revised:
2024-09-05
Accepted:
2024-10-09
Online:
2025-07-10
Published:
2025-07-10
Contact:
Yulin HE
About author:
CHEN Jiaqi, born in 1999, Ph. D. candidate. Her research interests include multi-sample statistical analysis, data mining, machine learning.Supported by:
摘要:
期望最大化(EM)算法在混合模型参数估计中发挥着重要作用,然而现有的EM算法在求解Gamma混合模型(GaMM)参数时存在局限性,主要体现在因近似计算导致的低质量参数估计,以及由于大量数值计算造成的计算效率低下问题。为了克服这些局限,并充分利用数据的多模性质,提出一种半EM(Semi-EM)算法求解用于估计多模概率分布的GaMM。首先,通过聚类探测数据的空间分布特性,以初始化GaMM参数,进而更准确地刻画数据的多模性;其次,在EM算法框架的基础上,对于缺乏封闭更新表达式而导致的参数更新困难问题,采用自定义的启发式策略对GaMM形状参数进行更新,使它们朝着最大化对数似然值的方向逐步调整,同时以封闭形式更新其他参数。经过一系列具有说服力的实验,验证了Semi-EM算法的可行性、合理性和有效性。实验结果表明,Semi-EM算法在精确估计多模概率分布方面优于对比的4种算法,具有更低的误差指标以及更高的对数似然值,表明该算法能提供更准确的模型参数估计,从而更精确地刻画数据的多模性质。
中图分类号:
陈佳琪, 何玉林, 成英超, 黄哲学. 求解多模概率分布Gamma混合模型的半EM算法[J]. 计算机应用, 2025, 45(7): 2153-2161.
Jiaqi CHEN, Yulin HE, Yingchao CHENG, Zhexue HUANG. Semi-EM algorithm for solving Gamma mixture model of multimodal probability distribution[J]. Journal of Computer Applications, 2025, 45(7): 2153-2161.
数据集 | 样本数 | 模数 | 权重 | 形状参数 | 尺度参数 | |
---|---|---|---|---|---|---|
训练集 | 测试集 | |||||
#1 | 1 000 | 2 000 | 3 | [0.33,0.50,0.17] | [ | [1.62,1.38,1.69] |
#2 | 1 000 | 2 000 | 4 | [0.24,0.24,0.24,0.28] | [90,25,54,73] | [2.55,1.37,2.53,1.23] |
#3 | 2 000 | 4 000 | 5 | [0.12,0.27,0.33,0.13,0.15] | [ | [1.37,1.39,1.43,1.43,1.39] |
#4 | 2 000 | 4 000 | 5 | [0.20,0.20,0.20,0.20,0.20] | [41,98,156,214,270] | [1.32,1.68,2.31,2.85,2.89] |
#5 | 2 500 | 5 000 | 6 | [0.09,0.08,0.10,0.38,0.14,0.21] | [ | [1.77,1.14,1.24,1.44,1.46,1.64] |
#6 | 2 500 | 5 000 | 6 | [0.16,0.16,0.16,0.16,0.16,0.20] | [48,102,153,210,262,318] | [1.96,2.56,2.50,2.33,2.94,2.92] |
#7 | 3 000 | 6 000 | 7 | [0.14,0.14,0.14,0.14,0.14,0.14,0.16] | [284,136,532,400,96,4,418] | [1.52,1.72,1.91,1.59,1.80,1.77,1.86] |
#8 | 3 000 | 6 000 | 8 | [0.12,0.12,0.12,0.12,0.12,0.12,0.12,0.16] | [ | [1.85,1.54,1.97,1.82,1.87,1.80,1.87,1.99] |
表1 实验中使用的8个仿真数据集
Tab. 1 Eight synthetic datasets used in experiments
数据集 | 样本数 | 模数 | 权重 | 形状参数 | 尺度参数 | |
---|---|---|---|---|---|---|
训练集 | 测试集 | |||||
#1 | 1 000 | 2 000 | 3 | [0.33,0.50,0.17] | [ | [1.62,1.38,1.69] |
#2 | 1 000 | 2 000 | 4 | [0.24,0.24,0.24,0.28] | [90,25,54,73] | [2.55,1.37,2.53,1.23] |
#3 | 2 000 | 4 000 | 5 | [0.12,0.27,0.33,0.13,0.15] | [ | [1.37,1.39,1.43,1.43,1.39] |
#4 | 2 000 | 4 000 | 5 | [0.20,0.20,0.20,0.20,0.20] | [41,98,156,214,270] | [1.32,1.68,2.31,2.85,2.89] |
#5 | 2 500 | 5 000 | 6 | [0.09,0.08,0.10,0.38,0.14,0.21] | [ | [1.77,1.14,1.24,1.44,1.46,1.64] |
#6 | 2 500 | 5 000 | 6 | [0.16,0.16,0.16,0.16,0.16,0.20] | [48,102,153,210,262,318] | [1.96,2.56,2.50,2.33,2.94,2.92] |
#7 | 3 000 | 6 000 | 7 | [0.14,0.14,0.14,0.14,0.14,0.14,0.16] | [284,136,532,400,96,4,418] | [1.52,1.72,1.91,1.59,1.80,1.77,1.86] |
#8 | 3 000 | 6 000 | 8 | [0.12,0.12,0.12,0.12,0.12,0.12,0.12,0.16] | [ | [1.85,1.54,1.97,1.82,1.87,1.80,1.87,1.99] |
数据集 | 属性 | 实例数 | 下载地址 |
---|---|---|---|
Concrete Compressive Strength | Mpa | 1 030 | https://archive.ics.uci.edu/dataset/165/concrete+compressive+strength |
HCV Data | AST | 589 | https://archive.ics.uci.edu/dataset/571/hcv+data |
Ozone Level Detection | WSR16 | 1 846 | https://archive.ics.uci.edu/dataset/172/ozone+level+detection |
Glass | RI | 213 | https://archive.ics.uci.edu/dataset/42/glass+identification |
表2 实验中使用的4个UCI据集
Tab. 2 Four UCI datasets used in experiments
数据集 | 属性 | 实例数 | 下载地址 |
---|---|---|---|
Concrete Compressive Strength | Mpa | 1 030 | https://archive.ics.uci.edu/dataset/165/concrete+compressive+strength |
HCV Data | AST | 589 | https://archive.ics.uci.edu/dataset/571/hcv+data |
Ozone Level Detection | WSR16 | 1 846 | https://archive.ics.uci.edu/dataset/172/ozone+level+detection |
Glass | RI | 213 | https://archive.ics.uci.edu/dataset/42/glass+identification |
算法 | 数据集#1 | 数据集#2 | ||||||||
---|---|---|---|---|---|---|---|---|---|---|
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.288 7E-06 | 0.001 355 86 | 3.301 7E-07 | 0.003 696 67 | -7 097.856 3 | 5.743 5E-07 | 0.002 148 45 | 2.538 5E-07 | 0.003 207 55 | -5 336.178 1 |
GaMM2006 | 2.199 7E-08 | 2.566 3E-05 | 6.610 6E-09 | 4.592 8E-05 | -7 077.843 4 | 1.731 7E-07 | 0.000 786 90 | 9.714 8E-08 | 0.001 563 3 | -5 332.252 7 |
GaMM2014 | 1.938 2E-08 | 2.406 2E-05 | 5.923 1E-09 | 4.232 9E-05 | -7 077.881 0 | 1.790 3E-06 | 0.008 408 44 | 1.086 8E-06 | 0.017 134 4 | -5 358.035 7 |
GaMM2019 | 2.257 4E-08 | 2.627 4E-05 | 6.797 E-09 | 4.775 4E-05 | -7 077.843 1 | 5.594 5E-06 | 0.030 393 91 | 3.548 3E-06 | 0.060 513 38 | -5 473.822 9 |
Semi-EM | 1.938 2E-08 | 2.406 3E-05 | 5.923 1E-09 | 4.232 9E-05 | -7 077.881 0 | 1.652 E-07 | 0.000 678 26 | 8.320 6E-08 | 0.001 218 64 | -5 332.837 8 |
算法 | 数据集#3 | 数据集#4 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.046 7E-07 | 0.000 820 69 | 6.086 5E-08 | 0.001 906 74 | -11 069.391 | 9.545 1E-08 | 0.001 262 04 | 2.532 9E-08 | 0.002 626 14 | -12 266.562 |
GaMM2006 | 8.077 0E-08 | 0.000 613 11 | 4.501 1E-08 | 0.001 078 25 | -11 069.841 | 1.578 6E-07 | 0.007 147 08 | 1.444 1E-07 | 0.021 879 93 | -12 389.272 |
GaMM2014 | 2.005 7E-07 | 0.001 547 10 | 1.108 1E-07 | 0.003 740 39 | -11 095.797 | 3.668 7E-08 | 0.000 753 73 | 1.524 3E-08 | 0.002 401 10 | -12 267.118 |
GaMM2019 | 2.741 7E-07 | 0.004 353 55 | 3.213 3E-07 | 0.010 848 10 | -11 127.867 | 1.575 6E-07 | 0.007 140 97 | 1.442 6E-07 | 0.021 870 18 | -12 389.271 |
Semi-EM | 7.232 4E-08 | 0.000 569 31 | 4.439 4E-08 | 0.001 375 90 | -11 074.342 | 2.623 E-08 | 0.000 544 58 | 1.178 2E-08 | 0.001 882 12 | -12 265.364 |
算法 | 数据集#5 | 数据集#6 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.096 5E-07 | 0.001 304 30 | 4.527 5E-08 | 0.002 518 51 | -14 724.125 | 1.549 1E-08 | 0.000 736 18 | 7.018 7E-09 | 0.001 360 99 | -16 355.412 |
GaMM2006 | 3.682 9E-07 | 0.009 209 18 | 3.378 6E-07 | 0.030 770 04 | -14 924.537 | 1.111 6E-08 | 0.000 435 45 | 4.730 6E-09 | 0.000 809 38 | -16 344.487 |
GaMM2014 | 1.223 8E-07 | 0.001 992 63 | 7.087 3E-08 | 0.005 347 13 | -14 761.696 | 1.913 1E-08 | 0.000 976 98 | 1.061 6E-08 | 0.002 099 25 | -16 352.326 |
GaMM2019 | 2.443 7E-08 | 0.000 420 71 | 1.474 E-08 | 0.001 004 11 | -14 721.810 | 6.435 4E-08 | 0.005 457 10 | 5.254 3E-08 | 0.010 626 59 | -16 409.141 |
Semi-EM | 4.946 5E-08 | 0.000 539 35 | 2.122 3E-08 | 0.001 160 53 | -14 724.010 | 1.116 3E-08 | 0.000 458 13 | 5.028 2E-09 | 0.000 855 10 | -16 344.724 |
算法 | 数据集#7 | 数据集#8 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.594 5E-06 | 0.007 736 24 | 1.108 4E-07 | 0.004 702 12 | -19 023.714 | 9.971 7E-08 | 0.002 164 32 | 2.631 9E-08 | 0.002 423 68 | -19 405.375 |
GaMM2006 | 3.573 1E-07 | 0.014 151 94 | 2.225 5E-07 | 0.063 368 13 | -19 543.963 | 1.966 2E-06 | 0.041 967 65 | 6.089 3E-07 | 0.065 312 87 | -19 930.687 |
GaMM2014 | 1.636 1E-06 | 0.007 726 38 | 1.548 9E-07 | 0.007 352 26 | -19 030.878 | 7.123 5E-08 | 0.002 724 77 | 3.277 5E-08 | 0.005 498 42 | -19 433.913 |
GaMM2019 | 2.091 8E-07 | 0.004 878 20 | 7.528 1E-08 | 0.009 700 39 | -19 077.768 | 6.554 6E-08 | 0.004 967 89 | 6.128 5E-08 | 0.010 790 24 | -19 495.266 |
Semi-EM | 5.500 2E-08 | 0.000 820 56 | 1.210 1E-08 | 0.001 840 06 | -18 986.337 | 1.908 1E-08 | 0.001 012 99 | 1.335 6E-08 | 0.002 153 18 | -19 402.666 |
表3 在仿真数据集上比较不同算法估计的PDF
Tab. 3 Comparison of PDFs estimated by using different algorithms on synthetic datasets
算法 | 数据集#1 | 数据集#2 | ||||||||
---|---|---|---|---|---|---|---|---|---|---|
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.288 7E-06 | 0.001 355 86 | 3.301 7E-07 | 0.003 696 67 | -7 097.856 3 | 5.743 5E-07 | 0.002 148 45 | 2.538 5E-07 | 0.003 207 55 | -5 336.178 1 |
GaMM2006 | 2.199 7E-08 | 2.566 3E-05 | 6.610 6E-09 | 4.592 8E-05 | -7 077.843 4 | 1.731 7E-07 | 0.000 786 90 | 9.714 8E-08 | 0.001 563 3 | -5 332.252 7 |
GaMM2014 | 1.938 2E-08 | 2.406 2E-05 | 5.923 1E-09 | 4.232 9E-05 | -7 077.881 0 | 1.790 3E-06 | 0.008 408 44 | 1.086 8E-06 | 0.017 134 4 | -5 358.035 7 |
GaMM2019 | 2.257 4E-08 | 2.627 4E-05 | 6.797 E-09 | 4.775 4E-05 | -7 077.843 1 | 5.594 5E-06 | 0.030 393 91 | 3.548 3E-06 | 0.060 513 38 | -5 473.822 9 |
Semi-EM | 1.938 2E-08 | 2.406 3E-05 | 5.923 1E-09 | 4.232 9E-05 | -7 077.881 0 | 1.652 E-07 | 0.000 678 26 | 8.320 6E-08 | 0.001 218 64 | -5 332.837 8 |
算法 | 数据集#3 | 数据集#4 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.046 7E-07 | 0.000 820 69 | 6.086 5E-08 | 0.001 906 74 | -11 069.391 | 9.545 1E-08 | 0.001 262 04 | 2.532 9E-08 | 0.002 626 14 | -12 266.562 |
GaMM2006 | 8.077 0E-08 | 0.000 613 11 | 4.501 1E-08 | 0.001 078 25 | -11 069.841 | 1.578 6E-07 | 0.007 147 08 | 1.444 1E-07 | 0.021 879 93 | -12 389.272 |
GaMM2014 | 2.005 7E-07 | 0.001 547 10 | 1.108 1E-07 | 0.003 740 39 | -11 095.797 | 3.668 7E-08 | 0.000 753 73 | 1.524 3E-08 | 0.002 401 10 | -12 267.118 |
GaMM2019 | 2.741 7E-07 | 0.004 353 55 | 3.213 3E-07 | 0.010 848 10 | -11 127.867 | 1.575 6E-07 | 0.007 140 97 | 1.442 6E-07 | 0.021 870 18 | -12 389.271 |
Semi-EM | 7.232 4E-08 | 0.000 569 31 | 4.439 4E-08 | 0.001 375 90 | -11 074.342 | 2.623 E-08 | 0.000 544 58 | 1.178 2E-08 | 0.001 882 12 | -12 265.364 |
算法 | 数据集#5 | 数据集#6 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.096 5E-07 | 0.001 304 30 | 4.527 5E-08 | 0.002 518 51 | -14 724.125 | 1.549 1E-08 | 0.000 736 18 | 7.018 7E-09 | 0.001 360 99 | -16 355.412 |
GaMM2006 | 3.682 9E-07 | 0.009 209 18 | 3.378 6E-07 | 0.030 770 04 | -14 924.537 | 1.111 6E-08 | 0.000 435 45 | 4.730 6E-09 | 0.000 809 38 | -16 344.487 |
GaMM2014 | 1.223 8E-07 | 0.001 992 63 | 7.087 3E-08 | 0.005 347 13 | -14 761.696 | 1.913 1E-08 | 0.000 976 98 | 1.061 6E-08 | 0.002 099 25 | -16 352.326 |
GaMM2019 | 2.443 7E-08 | 0.000 420 71 | 1.474 E-08 | 0.001 004 11 | -14 721.810 | 6.435 4E-08 | 0.005 457 10 | 5.254 3E-08 | 0.010 626 59 | -16 409.141 |
Semi-EM | 4.946 5E-08 | 0.000 539 35 | 2.122 3E-08 | 0.001 160 53 | -14 724.010 | 1.116 3E-08 | 0.000 458 13 | 5.028 2E-09 | 0.000 855 10 | -16 344.724 |
算法 | 数据集#7 | 数据集#8 | ||||||||
TMSE | TJSD | GMSE | GJSD | log-likelihood | TMSE | TJSD | GMSE | GJSD | log-likelihood | |
EM-GMM | 1.594 5E-06 | 0.007 736 24 | 1.108 4E-07 | 0.004 702 12 | -19 023.714 | 9.971 7E-08 | 0.002 164 32 | 2.631 9E-08 | 0.002 423 68 | -19 405.375 |
GaMM2006 | 3.573 1E-07 | 0.014 151 94 | 2.225 5E-07 | 0.063 368 13 | -19 543.963 | 1.966 2E-06 | 0.041 967 65 | 6.089 3E-07 | 0.065 312 87 | -19 930.687 |
GaMM2014 | 1.636 1E-06 | 0.007 726 38 | 1.548 9E-07 | 0.007 352 26 | -19 030.878 | 7.123 5E-08 | 0.002 724 77 | 3.277 5E-08 | 0.005 498 42 | -19 433.913 |
GaMM2019 | 2.091 8E-07 | 0.004 878 20 | 7.528 1E-08 | 0.009 700 39 | -19 077.768 | 6.554 6E-08 | 0.004 967 89 | 6.128 5E-08 | 0.010 790 24 | -19 495.266 |
Semi-EM | 5.500 2E-08 | 0.000 820 56 | 1.210 1E-08 | 0.001 840 06 | -18 986.337 | 1.908 1E-08 | 0.001 012 99 | 1.335 6E-08 | 0.002 153 18 | -19 402.666 |
算法 | Concrete Compressive Strength | HCV Data | Ozone Level Detection | Glass |
---|---|---|---|---|
EM-GMM | -4 323.56 | -2 226.41 | -3 053.25 | 549.012 1 |
GaMM2006 | -4 314.74 | -2 211.78 | -3 037.53 | 555.751 1 |
GaMM2014 | -4 329.64 | -2 256.15 | -3 064.82 | 492.182 3 |
GaMM2019 | -4 320.15 | -2 217.24 | -3 037.53 | 540.282 1 |
Semi-EM | -4 314.74 | -2 209.43 | -3 037.52 | 553.408 5 |
表4 在UCI数据集上比较不同算法得到的对数似然值
Tab. 4 Comparison of log-likelihood values estimated by using different algorithms on UCI datasets
算法 | Concrete Compressive Strength | HCV Data | Ozone Level Detection | Glass |
---|---|---|---|---|
EM-GMM | -4 323.56 | -2 226.41 | -3 053.25 | 549.012 1 |
GaMM2006 | -4 314.74 | -2 211.78 | -3 037.53 | 555.751 1 |
GaMM2014 | -4 329.64 | -2 256.15 | -3 064.82 | 492.182 3 |
GaMM2019 | -4 320.15 | -2 217.24 | -3 037.53 | 540.282 1 |
Semi-EM | -4 314.74 | -2 209.43 | -3 037.52 | 553.408 5 |
[1] | McLACHLAN G J, LEE S X, RATHNAYAKE S I. Finite mixture models [J]. Annual Review of Statistics and Its Application, 2019, 6: 355-378. |
[2] | AN P, WANG Z, ZHANG C. Ensemble unsupervised autoencoders and Gaussian mixture model for cyberattack detection [J]. Information Processing and Management, 2022, 59(2): No.102844. |
[3] | YANG B, LIU C, LI B, et al. Prototype mixture models for few-shot semantic segmentation [C]// Proceedings of the 2020 European Conference on Computer Vision, LNCS 12353. Cham: Springer, 2020: 763-778. |
[4] | ZHU Y, TANG Y, TANG Y, et al. Cross-domain medical image translation by shared latent Gaussian mixture model [C]// Proceedings of the 2020 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 12262. Cham: Springer, 2020: 379-389. |
[5] | 朱福敏,宋佳音,刘仪榕.基于Levy-GARCH模型的股票市场尾部风险度量研究[J].中央财经大学学报,2024(1): 47-60. |
ZHU F M, SONG J Y, LIU Y R. Tail risk measurement in stock market based on Levy-GARCH model [J]. Journal of Central University of Finance and Economics, 2024(1): 47-60. | |
[6] | BURGESS-HULL A J. Finite mixture models with student t distributions: an applied example [J]. Prevention Science, 2020, 21(6): 872-883. |
[7] | 段明义,李祖照,崔奥杰.基于改进t分布混合模型的路面裂缝图像分割方法研究[J].公路交通科技,2022, 39(7): 23-29. |
DUAN M Y, LI Z Z, CUI A J. Study on segmentation method of pavement crack image based on improved t-distribution mixture model [J]. Journal of Highway and Transportation Research and Development, 2022, 39(7): 23-29. | |
[8] | WANG Q, GUO G, QIAN G, et al. Distributed online expectation-maximization algorithm for Poisson mixture model [J]. Applied Mathematical Modelling, 2023, 124: 734-748. |
[9] | DANG U J, BROWNE R P, McNICHOLAS P D. Mixtures of multivariate power exponential distributions [J]. Biometrics, 2015, 71(4): 1081-1089. |
[10] | BROWNE R P, McNICHOLAS P D. A mixture of generalized hyperbolic distributions [J]. Canadian Journal of Statistics, 2015, 43(2): 176-198. |
[11] | LIU C, LI H C, FU K, et al. Bayesian estimation of generalized Gamma mixture model based on variational EM algorithm [J]. Pattern Recognition, 2019, 87: 269-284. |
[12] | KOSMIDIS I, KARLIS D. Model-based clustering using copulas with applications [J]. Statistics and Computing, 2016, 26(5): 1079-1099. |
[13] | MARBAC M, BIERNACKI C, VANDEWALLE V. Model-based clustering of Gaussian copulas for mixed data [J]. Communications in Statistics — Theory and Methods, 2017, 46(23): 11635-11656. |
[14] | SAHIN Ö, CZADO C. Vine copula mixture models and clustering for non-Gaussian data [J]. Econometrics and Statistics, 2022, 22: 136-158. |
[15] | ALMHANA J, LIU Z, CHOULAKIAN V, et al. A recursive algorithm for Gamma mixture models [C]// Proceedings of the 2006 IEEE International Conference on Communications — Volume 1. Piscataway: IEEE, 2006: 197-202. |
[16] | LI Y, CHUN Y J. Stochastic geometric analysis of IRS-aided wireless networks using mixture Gamma model [C]// Proceedings of the 2021 International Conference on Innovative Mobile and Internet Services in Ubiquitous Computing, LNNS 279. Cham: Springer, 2022: 168-178. |
[17] | SHI X, WANG Y, YOU H, et al. Sea ice extraction in SAR images via a spatially constrained Gamma mixture model [J]. Sustainability, 2023, 15(13): No.10374. |
[18] | JIA D W, WU Z Y. Structural reliability analysis under stochastic seismic excitations and multidimensional limit state based on Gamma mixture model and copula function [J]. Probabilistic Engineering Mechanics, 2024, 76: No.103621. |
[19] | AZHAR M, LI M J, HUANG J Z. A hierarchical Gamma mixture model-based method for classification of high-dimensional data [J]. Entropy, 2019, 21(9): No.906. |
[20] | SALLY H, BOUROUIS S, BOUGUILA N. Online learning of finite and infinite Gamma mixture models for COVID-19 detection in medical images [J]. Computers, 2020, 10(1): No.6. |
[21] | VEGAS-SÁNCHEZ-FERRERO G, MARTÍN-FERNÁNDEZ M, SANCHES J M. A Gamma mixture model for IVUS imaging [M]// SABA L, SANCHES J M, PEDRO L M, et al. Multi-modality atherosclerosis imaging and diagnosis. New York: Springer, 2014: 155-171. |
[22] | BENLAKHDAR S, RZIZA M, THAMI R O H. A robust model using SIFT and Gamma mixture model for texture images classification: perspectives for medical applications [J]. Biomedical and Pharmacology Journal, 2020, 13(4): 1659-1669. |
[23] | 魏倩,何冰冰,张榆锋,等.基于伽马混合模型B超图像聚类配准的颈动脉管壁搏动位移干扰抑制[J].仪器仪表学报,2022, 43(7): 199-208. |
WEI Q, HE B B, ZHANG Y F, et al. Carotid artery wall pulsation displacement interference suppression based on ultrasonic images clustered registration with the Gamma mixture model [J]. Chinese Journal of Scientific Instrument, 2022, 43(7): 199-208. | |
[24] | DEMPSTER A P, LAIRD N M, RUBIN D B. Maximum likelihood from incomplete data via the EM algorithm [J]. Journal of the Royal Statistical Society: Series B (Methodological), 1977, 39(1): 1-22. |
[25] | WANG J, HANCOCK E R. Statistical structural inference from edge weights using a mixture of Gamma distributions [J]. Journal of Complex Networks, 2023, 11(5): No.cnad038. |
[26] | YOUNG D S, CHEN X, HEWAGE D C, et al. Finite mixture-of-Gamma distributions: estimation, inference, and model-based clustering [J]. Advances in Data Analysis and Classification, 2019, 13(4): 1053-1082. |
[27] | HE M, CHEN J. Consistency of the MLE under a two-parameter Gamma mixture model with a structural shape parameter [J]. Metrika, 2022, 85(8): 951-975. |
[28] | HE M, CHEN J. Strong consistency of the MLE under two-parameter Gamma mixture models with a structural scale parameter [J]. Advances in Data Analysis and Classification, 2022, 16(1): 125-154. |
[29] | NAIDOO B, XU H. Maximum likelihood estimation of the Gamma and Gamma mixture distribution [C]// Proceedings of the 2023 IEEE AFRICON Conference. Piscataway: IEEE, 2023: 1-6. |
[30] | ZOU Y H, LI H C. MCMC estimation of finite generalized Gamma mixture model [C]// Proceedings of the 2012 IEEE International Geoscience and Remote Sensing Symposium. Piscataway: IEEE, 2012: 5935-5938. |
[31] | SUSANTO I, IRIAWAN N, KUSWANTO H, et al. Bayesian inference for the finite Gamma mixture model of income distribution [J]. Journal of Physics: Conference Series, 2019, 1217: No.012077. |
[32] | LAI Y, CAO H, LUO L, et al. Extended variational inference for Gamma mixture model in positive vectors modeling [J]. Neurocomputing, 2021, 432: 145-158. |
[33] | BOUROUIS S, BOUGUILA N. Expectation propagation learning of finite and infinite Gamma mixture models and its applications [J]. Multimedia Tools and Applications, 2023, 82(21): 33267-33284. |
[34] | ARTHUR D, VASSILVITSKII S. k-means++: the advantages of careful seeding [C]// Proceedings of the 18th Annual ACM-SIAM symposium on Discrete algorithms. Philadelphia, PA: SIAM, 2007: 1027-1035. |
[35] | DUA D, GRAFF C. UCI machine learning repository [DB/OL]. [2024-01-25]. . |
[1] | 郭书君, 任卫军, 陈倩倩, 游广飞. 基于聚类多变量时间序列模型的交通状态实时预测[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2253-2261. |
[2] | 徐乐, 黄瑞章, 白瑞娜, 秦永彬. 基于意图正则化的深度半监督文本聚类[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2145-2152. |
[3] | 谷铮, 陈学斌, 张宏扬, 李雨欣. 基于凝聚式层次聚类的微调筛选过采样方法[J]. 《计算机应用》唯一官方网站, 2025, 45(7): 2138-2144. |
[4] | 王建华, 吴传宇, 许莉萍. 多因素柔性作业车间绿色调度的改进进化算法[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1954-1962. |
[5] | 崔双双, 王宏志, 朱加昊, 吴昊. 面向低能耗高性能的分类器两阶段数据选择方法[J]. 《计算机应用》唯一官方网站, 2025, 45(6): 1703-1711. |
[6] | 罗蒙, 高超, 王震. 基于带约束谱聚类的启发式车辆路径规划算法优化方法[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1387-1394. |
[7] | 郭书剑, 余节约, 尹学松. 图正则化弹性网子空间聚类[J]. 《计算机应用》唯一官方网站, 2025, 45(5): 1464-1471. |
[8] | 陈庆礼, 郭渊博, 方晨. 面向数据异构的聚类联邦学习算法[J]. 《计算机应用》唯一官方网站, 2025, 45(4): 1086-1094. |
[9] | 区卓越, 邓秀勤, 陈磊. 基于加权锚点的自适应多视图互补聚类算法[J]. 《计算机应用》唯一官方网站, 2025, 45(1): 115-126. |
[10] | 李顺勇, 李师毅, 胥瑞, 赵兴旺. 基于自注意力融合的不完整多视图聚类算法[J]. 《计算机应用》唯一官方网站, 2024, 44(9): 2696-2703. |
[11] | 王清, 赵杰煜, 叶绪伦, 王弄潇. 统一框架的增强深度子空间聚类方法[J]. 《计算机应用》唯一官方网站, 2024, 44(7): 1995-2003. |
[12] | 董瑶, 付怡雪, 董永峰, 史进, 陈晨. 不完整多视图聚类综述[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1673-1682. |
[13] | 蒋小霞, 黄瑞章, 白瑞娜, 任丽娜, 陈艳平. 基于事件表示和对比学习的深度事件聚类方法[J]. 《计算机应用》唯一官方网站, 2024, 44(6): 1734-1742. |
[14] | 黄天宇, 李远兴, 陈昊, 郭紫佳, 魏明军. 地空协同场景下加权模糊聚类用户簇划分方法[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1555-1561. |
[15] | 徐童童, 解滨, 张春昊, 张喜梅. 融合转移概率矩阵的多阶最近邻图聚类算法[J]. 《计算机应用》唯一官方网站, 2024, 44(5): 1527-1538. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||