Skip to main content
. Author manuscript; available in PMC: 2011 Feb 25.
Published in final edited form as: Stat Comput. 2011 Jan 4;21(2):261–273. doi: 10.1007/s11222-009-9166-3

Table 5.

Comparison of the various algorithms for estimating the location and scale of a 10-variate t distribution with 0.5 degrees of freedom. The column ln L lists the average converged log-likelihood, and the column Evals lists the average number of EM evaluations. Running times are averaged over 100 simulations with 200 sample points each. The number of parameters is 65, and the stopping criterion is 10−9

D.F. Method EM
PX-EM
ln L Evals Time ln L Evals Time
1 EM −3981.5470 160 0.8272 −3981.5470 15 0.0771
q = 1 −3981.5470 26 0.1363 −3981.5470 10 0.0497
q = 2 −3981.5470 22 0.1184 −3981.5470 10 0.0510
q = 3 −3981.5470 23 0.1216 −3981.5470 10 0.0540
q = 4 −3981.5470 24 0.1282 −3981.5470 11 0.0555
q = 5 −3981.5470 26 0.1381 −3981.5470 11 0.0558
SqS1 −3981.5470 29 0.1570 −3981.5470 10 0.0509
SqS2 −3981.5470 31 0.1646 −3981.5470 10 0.0507
SqS3 −3981.5470 30 0.1588 −3981.5470 10 0.0507
0.5 EM −3975.8332 259 1.3231 −3975.8332 15 0.0763
q = 1 −3975.8332 31 0.1641 −3975.8332 10 0.0506
q = 2 −3975.8332 25 0.1343 −3975.8332 10 0.0512
q = 3 −3975.8332 27 0.1405 −3975.8332 10 0.0544
q = 4 −3975.8332 28 0.1479 −3975.8332 10 0.0547
q = 5 −3975.8332 30 0.1553 −3975.8332 11 0.0552
SqS1 −3975.8332 34 0.1829 −3975.8332 10 0.0514
SqS2 −3975.8332 38 0.2017 −3975.8332 10 0.0513
SqS3 −3975.8332 35 0.1895 −3975.8332 10 0.0513
0.1 EM −4114.2561 899 4.5996 −4114.2561 16 0.0816
q = 1 −4114.2562 52 0.2709 −4114.2561 10 0.0521
q = 2 −4114.2561 36 0.1924 −4114.2561 10 0.0533
q = 3 −4114.2561 34 0.1820 −4114.2561 10 0.0544
q = 4 −4114.2561 36 0.1895 −4114.2561 10 0.0544
q = 5 −4114.2561 38 0.2041 −4114.2561 11 0.0558
SqS1 −4114.2561 51 0.2717 −4114.2561 10 0.0522
SqS2 −4114.2561 66 0.3492 −4114.2561 10 0.0518
SqS3 −4114.2561 54 0.2846 −4114.2561 10 0.0519
0.05 EM −4224.9190 1596 8.1335 −4224.9190 17 0.0857
q = 1 −4224.9192 62 0.3248 −4224.9190 10 0.0530
q = 2 −4224.9192 47 0.2459 −4224.9190 10 0.0539
q = 3 −4224.9191 39 0.2006 −4224.9190 10 0.0549
q = 4 −4224.9191 40 0.2089 −4224.9190 11 0.0564
q = 5 −4224.9191 42 0.2239 −4224.9190 11 0.0565
SqS1 −4224.9191 60 0.3156 −4224.9190 10 0.0543
SqS2 −4224.9191 91 0.4809 −4224.9190 10 0.0535
SqS3 −4224.9191 64 0.3417 −4224.9190 10 0.0535