我使用df 8从8维t-copula进行模拟，所有维度之间的相关性为0.1。
有3种不同的边缘类型，以下边缘在copula中出现两次：
以下分布在8维copula中出现了四次：
我指定了所有的参数，定义了mvdc并用不同的种子数画了50次随机变量，我总能找到这种奇怪的模式，其中col 1＆amp; 2在奇数行中总是很高而在偶数行中总是很低。相关的copula函数如下：
myCop.t <- ellipCopula(family = "t", dim = dim, dispstr = "toep", param = copcor, df = df)
multi.distr <- mvdc(myCop.t, distrvec, paravec, marginsIdentical = FALSE,
check = FALSE, fixupNames = TRUE)
randoms <- rMvdc(50, multi.distr)
copula函数中的各个参数提供了正确的参数，因此在这里不感兴趣，我强烈怀疑实际的代码根本不感兴趣。
数据输出如下所示：
V1 V2 V3 V4 V5 V6 V7 V8
1 16.010861 14.996232 5.701437 6.787819 5.475126 6.990366 1.69338912 2.666378900
2 7.014132 5.563823 6.515101 4.660104 5.957121 6.781011 0.86962636 2.034051794
3 15.672679 15.749123 6.239703 5.087319 5.661866 5.419313 -0.03725051 0.115770530
4 4.323271 7.001562 5.941222 6.291748 4.793068 5.752201 2.03728897 1.857472281
5 14.368017 16.878522 4.601959 4.832871 4.713307 5.335259 -0.65059333 -0.006805472
6 6.407364 4.536828 5.434258 5.716145 5.054762 5.886708 0.76146941 0.662944596
7 15.323959 13.996845 4.740232 6.051559 3.994886 3.556641 -1.00196874 1.034595276
8 3.739731 5.284797 5.137518 5.345029 4.224926 5.133985 0.60952885 0.232399508
9 16.092425 15.895005 6.255278 6.480924 5.283971 4.965157 0.97320818 0.516121522
10 4.333453 5.208740 5.074339 5.606373 4.489304 5.094264 1.04653615 -0.055464950
11 15.211897 16.096414 6.372798 5.259175 6.041479 4.722053 -0.30172391 -0.199618446
12 5.851663 4.997275 6.050953 4.413398 5.323346 5.326618 1.07938868 1.115312951
13 15.815942 15.572386 5.463940 6.088293 5.670306 5.300775 1.56742392 0.686777736
14 4.446224 4.163309 4.181659 4.777350 4.568136 4.213055 -0.72199878 -0.824398248
15 14.785817 16.274832 5.669137 5.589509 4.655658 5.689984 0.19452841 1.349848277
16 4.109499 4.473835 3.919780 3.214149 4.263928 5.531912 -1.11236543 -1.303189568
17 14.239574 13.276767 4.200647 4.781725 4.326090 4.338101 0.70853038 -0.156449752
18 6.898483 6.994064 7.815153 6.462504 5.877854 5.384026 1.29072261 0.293196041
19 15.412930 15.349961 5.018926 5.922361 5.125757 6.700596 -1.12301994 -0.399204202
20 5.079742 4.748265 5.457855 4.654723 4.731529 4.188034 -0.34238254 -0.690145740
21 14.141139 13.888448 3.855420 4.827962 4.275877 3.921048 -0.70280816 -1.123617466
22 5.379309 6.428976 6.218161 4.924798 6.039379 5.474278 -0.05402753 0.808884618
23 15.810009 16.474434 4.722685 6.654449 5.799382 5.148503 0.21074129 -2.691441462
24 4.504196 4.134488 6.133392 5.341009 4.146695 5.485827 -0.17429215 0.377224880
25 15.090372 18.562274 6.940438 8.702905 7.421064 7.234131 0.95681780 13.957746201
26 5.169069 4.689696 3.945228 3.831344 4.257807 4.485700 -1.25036470 -1.015452283
27 14.231269 14.573801 3.979783 3.975574 4.059549 3.429011 -0.84782089 -0.708102232
28 3.466399 4.809140 4.642554 4.000181 5.514467 3.702483 -0.28220028 -2.230534184
29 14.535945 15.068252 5.389390 3.925671 4.820220 5.385509 0.44983724 0.369502404
30 5.721859 6.277543 4.951993 6.146136 5.522600 5.546753 2.97382789 -0.285541576
31 14.032321 14.633016 6.217122 3.954871 5.537166 6.034675 1.20169471 -0.023094229
32 4.866420 4.272465 3.896341 5.120164 3.583863 4.909106 -0.06676720 -0.296015677
33 14.833177 15.180053 4.989203 4.202957 4.889351 5.098410 -3.42600028 0.538323743
34 4.611121 4.268716 4.238652 3.660678 3.547858 4.580510 -1.77435703 -1.665450739
35 14.528331 13.893190 5.167363 4.361532 4.512446 4.611733 -1.85922632 -0.082607647
36 4.388169 4.139480 5.073937 5.257833 4.617510 4.309536 -0.97743122 -0.813389458
37 16.048187 17.595425 6.047185 5.723189 5.221158 5.160128 0.83381764 0.479124155
38 6.132494 4.633051 5.682781 5.747092 5.825939 5.104212 -0.07316043 1.717729460
39 15.559974 14.191118 6.874427 6.107418 6.731673 5.141445 0.10187992 0.351784075
40 5.126939 4.776856 5.626305 5.688585 6.054434 4.635697 0.68777501 0.653913128
41 14.687924 13.981411 5.854028 3.827776 3.636622 4.875557 -3.74749849 -1.843810747
42 4.632315 4.789424 5.133586 4.137874 5.002943 4.692032 0.67748217 -0.410707115
43 15.128650 16.278525 4.222412 3.912493 4.496541 3.582153 2.61045652 0.319884730
44 3.808546 5.055248 4.049429 4.836402 4.626999 4.031213 -1.70918390 -0.074491627
45 13.928535 13.631071 3.207690 3.501214 3.869997 3.519227 -2.32473887 -0.374699516
46 5.291424 5.473507 5.206400 4.055007 5.063000 4.069546 0.88998427 -0.055496756
47 15.596255 15.095043 4.619923 5.180290 5.550825 5.968506 2.15353841 0.063891460
48 6.471105 5.759449 5.731538 5.544783 5.612262 5.214522 1.15215748 0.607186739
49 15.159832 15.716472 4.666025 5.146686 5.394190 5.999513 1.27856163 -0.191519256
50 3.826062 4.010362 4.285219 3.725003 4.648102 3.732760 -0.54854635 -0.721668612
真正奇怪的是，当我更改第1列和第2列中正态分布的参数时，正态分布的平均值仍为15，但sd也是15（之前为5）数据似乎是＆＃34;对＆＃34;再次，行之间没有模式。样本输出的模拟如下：
V1 V2 V3 V4 V5 V6 V7 V8
1 14.83778 15.50690 5.404255 6.529869 3.723583 5.804804 -0.023806934 1.31729448
2 14.89108 15.52742 4.546552 4.133864 3.798082 3.944486 -0.958503275 -0.25154357
3 15.18946 14.42179 3.942194 5.375967 5.513438 5.338239 -0.941209336 1.02601868
4 13.89376 15.06167 5.741713 4.913896 4.800117 6.500235 -0.221074057 -0.32954484
5 15.25007 14.58085 4.518882 5.137600 4.328629 5.634427 -0.934152880 -1.50506026
6 17.57606 16.87651 6.309456 4.743752 4.066471 5.642209 0.777167206 -0.22662543
7 17.08080 12.29113 6.133664 5.059785 5.264676 4.733923 -0.219459993 -1.12209908
8 15.93970 14.43690 4.057061 5.064182 5.326703 5.564106 -1.250717314 0.94948212
9 16.24507 14.48802 4.975028 5.761803 6.711075 4.221429 -1.024120847 -0.33025094
10 15.47580 14.91176 4.118495 4.555365 5.210769 5.204387 0.829185436 -0.04127958
11 13.50148 16.05232 5.492861 4.048113 5.125046 6.086061 0.167497301 -1.17138922
12 14.30186 15.19833 5.076858 3.775417 5.300832 5.992216 0.164344390 0.68483847
13 15.91798 14.47786 3.007620 6.224872 3.830338 3.028243 -0.071242353 0.81984925
14 17.41911 14.49935 5.013210 5.169754 5.544124 5.201758 0.338119344 -0.35646076
15 15.76726 13.03769 3.462312 5.094576 6.251835 5.644511 0.071842820 -2.48345545
16 14.99288 14.54641 5.733469 6.241158 6.111138 4.372109 -1.002280884 0.76748645
17 15.17871 14.88004 6.631951 4.846311 6.077040 4.385127 -0.699274238 -1.84938212
18 12.90336 15.18397 4.481494 6.204927 3.310052 2.046945 -1.842346138 -0.46791915
19 14.31310 14.32862 6.963597 5.853332 4.124151 4.093026 -1.662460882 -0.27031668
20 17.25331 16.73606 4.674062 5.949045 5.169811 4.856738 0.048013929 1.07306298
21 15.38818 13.76637 4.897191 5.626547 5.076134 5.823196 0.137200385 1.40611275
22 15.35279 16.89325 5.150686 4.490542 3.409013 5.184520 0.004938741 -1.45164671
23 15.88499 15.31334 5.260789 4.401654 4.452729 6.435039 0.789386209 -0.50842299
24 16.06672 14.82119 5.024210 4.587017 5.758393 5.662651 0.231251570 -0.12430524
25 14.93408 15.62573 3.730154 3.625504 4.775070 3.788278 0.830299184 0.97256207
26 16.20754 14.23117 4.843946 4.206059 3.778496 3.346761 1.614550837 -1.75024630
27 16.71283 11.65426 3.559865 4.701199 4.198006 6.891123 -0.454631494 -0.42152501
28 15.15844 13.94274 4.657491 6.611439 6.503309 5.580698 1.133942406 0.44039104
29 14.26423 15.83374 5.961393 5.223553 5.173916 4.709981 1.424424855 -0.43280752
30 15.67209 15.29075 5.740236 4.233706 5.582893 4.706259 1.572087272 0.38300052
31 14.21829 14.83687 4.442544 5.778498 4.822692 6.121222 -0.318373779 -0.27456086
32 14.98008 14.26653 5.844600 4.832860 4.213410 5.809996 -1.779622235 0.07399854
33 15.63392 15.19140 5.263267 5.305941 5.929871 5.175858 0.086394088 -0.34753023
34 13.75393 14.16023 3.510472 5.739289 3.397409 5.158956 -0.378902314 -0.45783589
35 15.87641 13.69308 4.151437 5.504838 4.602269 4.988841 -0.618535468 -1.13646497
36 14.95559 15.79736 4.811952 5.113602 5.195094 4.568496 0.010490101 -1.01004477
37 13.83409 13.25046 6.686703 3.514009 5.308698 7.868867 9.791588382 2.05239012
38 15.60604 14.11555 6.362565 5.440548 5.503359 5.640982 -0.111587857 -0.39176421
39 15.96660 14.28885 5.726868 5.199140 5.990608 4.585727 1.768536285 -0.08786540
40 15.65997 16.64071 5.274348 5.406443 3.618235 7.037277 -2.381014003 0.38575367
41 14.18519 16.27092 4.780902 4.313851 5.706422 5.786148 0.024784912 -1.41593055
42 14.97552 12.97847 4.106621 5.258110 5.112121 5.313409 0.553056422 2.90330487
43 16.29216 16.30951 6.591961 5.021240 6.155636 3.628142 -1.787351964 -0.08872408
44 13.83075 14.44732 4.561066 5.177443 4.070881 5.955008 -1.050899776 0.89126680
45 14.44366 15.77155 6.323065 5.218095 3.199359 5.855317 -0.452537507 0.20151892
46 13.96975 15.67864 4.051353 5.602012 4.278524 4.870577 0.304695524 0.06143125
47 13.64975 13.30919 3.409978 7.107219 2.358819 5.178817 -0.486368335 -2.70001802
48 17.67098 15.15840 4.435062 6.742152 4.131794 5.817331 -1.161750029 0.31463652
49 15.62723 14.95017 5.157411 6.137096 4.993191 5.129493 -0.058337660 1.18703728
50 14.39389 13.07334 5.234347 3.257008 3.612170 4.721898 1.070309092 -0.59448545
那么这里发生了什么？在原始示例中，与其他两个边缘类型相比，前两列具有低标准偏差，而在第二个列中，所有正态分布具有与其平均值相同的标准。这是多维copula的固有特征，相对而言，边缘具有完全不同的波动率，或者可能是copula包的基本算法不准确？