where 
and 
correspond to dummy variables generated from classification variables in columns 5 and 6 of x
. Respectively, 
corresponds to column index 2, 
corresponds to column index 3, and 
corresponds to column index 4 of x
. Column 0 of x
contains the response and column 1 of x
contains the censoring code. Logging is used to print output statistics.
import com.imsl.stat.*;
import com.imsl.math.PrintMatrix;
import java.util.logging.*;
public class ProportionalHazardsEx1 {
public static void main(String args[]) throws Exception{
double[][] x = {
{411, 0, 7, 64, 5, 1, 0},
{126, 0, 6, 63, 9, 1, 0},
{118, 0, 7, 65, 11, 1, 0},
{ 92, 0, 4, 69, 10, 1, 0},
{ 8, 0, 4, 63, 58, 1, 0},
{ 25, 1, 7, 48, 9, 1, 0},
{ 11, 0, 7, 48, 11, 1, 0},
{ 54, 0, 8, 63, 4, 2, 0},
{153, 0, 6, 63, 14, 2, 0},
{ 16, 0, 3, 53, 4, 2, 0},
{ 56, 0, 8, 43, 12, 2, 0},
{ 21, 0, 4, 55, 2, 2, 0},
{287, 0, 6, 66, 25, 2, 0},
{ 10, 0, 4, 67, 23, 2, 0},
{ 8, 0, 2, 61, 19, 3, 0},
{ 12, 0, 5, 63, 4, 3, 0},
{177, 0, 5, 66, 16, 4, 0},
{ 12, 0, 4, 68, 12, 4, 0},
{200, 0, 8, 41, 12, 4, 0},
{250, 0, 7, 53, 8, 4, 0},
{100, 0, 6, 37, 13, 4, 0},
{999, 0, 9, 54, 12, 1, 1},
{231, 1, 5, 52, 8, 1, 1},
{991, 0, 7, 50, 7, 1, 1},
{ 1, 0, 2, 65, 21, 1, 1},
{201, 0, 8, 52, 28, 1, 1},
{ 44, 0, 6, 70, 13, 1, 1},
{ 15, 0, 5, 40, 13, 1, 1},
{103, 1, 7, 36, 22, 2, 1},
{ 2, 0, 4, 44, 36, 2, 1},
{ 20, 0, 3, 54, 9, 2, 1},
{ 51, 0, 3, 59, 87, 2, 1},
{ 18, 0, 4, 69, 5, 3, 1},
{ 90, 0, 6, 50, 22, 3, 1},
{ 84, 0, 8, 62, 4, 3, 1},
{164, 0, 7, 68, 15, 4, 1},
{ 19, 0, 3, 39, 4, 4, 1},
{ 43, 0, 6, 49, 11, 4, 1},
{340, 0, 8, 64, 10, 4, 1},
{231, 0, 7, 67, 18, 4, 1}
};
int indef[] = {2, 3, 4, 5, 6};
int nvef[] = {1, 1, 1, 1, 1};
int indcl[] = {5, 6};
int maxcl = 6, icen = 1;
double ratio = 10000.0;
ProportionalHazards ph = new ProportionalHazards(x, nvef, indef);
ph.setMaxClass(maxcl);
ph.setCensorColumn(icen);
ph.setClassVarColumns(indcl);
ph.setStratumRatio(ratio);
// Level.FINER prints most output statistics
Logger logger = ph.getLogger();
ConsoleHandler ch = new ConsoleHandler();
// Set ConsoleHandler Level to ALL
ch.setLevel(Level.ALL);
logger.setLevel(Level.FINER);
logger.addHandler(ch);
ch.setFormatter(new com.imsl.IMSLFormatter());
double coef[][] = ph.getParameterStatistics();
new PrintMatrix("\nFinal Coefficient Matrix").print(coef);
}
}
Initial Estimates
0
0 0
1 0
2 0
3 0
4 0
5 0
6 0
Method Iteration Step Maximum scaled Log
Size coef. update likelihood
Q-N 0 -102.40056551567265
Q-N 1 1.0 0.503384047154466 -91.04395077806367
Q-N 2 1.0 0.5781995573157527 -88.06808499091763
N-R 3 1.0 0.11310452893592865 -87.92233446871525
N-R 4 1.0 0.0695795111007413 -87.88778754387496
N-R 5 1.0 8.149048655014991E-4 -87.88778009939108
Log-Likelihood = -87.888
Coefficient Statistics
Coefficient Std. error Asymptotic z-stat Asymptotic p-value
0 -0.5846 0.1368 -4.2721 0
1 -0.0131 0.0206 -0.6342 0.526
2 0.0008 0.0118 0.0645 0.9486
3 -0.367 0.4848 -0.7572 0.449
4 -0.0077 0.5068 -0.0152 0.9878
5 1.1129 0.6331 1.758 0.0787
6 0.3797 0.4058 0.9357 0.3494
Asymptotic Coefficient Covariance
0 1 2 3 4 5 6
0 0.0187 0.0003 0.0003 0.0057 0.0097 0.0043 0.0021
1 0.0004 -0 -0.0017 -0.0008 -0.0031 -0.0029
2 0.0001 0.0008 -0.0018 0.0006 0.0017
3 0.235 0.098 0.1184 0.0373
4 0.2568 0.1253 -0.0194
5 0.4008 0.0629
6 0.1647
Case Analysis
Survival Prob. Influence Residual Cumulative hazard Prop. constant
0 0.0022 0.0414 2.0531 6.1032 0.3364
1 0.2988 0.1088 0.7409 1.2078 0.6134
2 0.3424 0.1184 0.3576 1.0719 0.3336
3 0.4336 0.1554 1.5272 0.8357 1.8274
4 0.9555 0.5567 0.0933 0.0455 2.0499
5 0.7365 ? 0.1272 0.3058 0.4158
6 0.9204 0.3729 0.0346 0.083 0.4164
7 0.5876 0.2637 0.1446 0.5317 0.2719
8 0.2577 0.1173 1.196 1.3561 0.882
9 0.8457 0.1486 0.9656 0.1676 5.7608
10 0.5481 0.3133 0.2135 0.6012 0.3551
11 0.7365 0.2108 0.9551 0.3058 3.1232
12 0.0293 0.0602 3.018 3.5289 0.8552
13 0.9382 0.0935 0.173 0.0638 2.7135
14 0.9555 0.1595 1.3142 0.0455 28.8855
15 0.8854 0.2322 0.5864 0.1217 4.8164
16 0.1814 0.0918 2.6217 1.707 1.5358
17 0.8854 0.1869 0.3258 0.1217 2.6765
18 0.1414 0.2303 0.7187 1.9565 0.3673
19 0.0522 0.0943 1.6591 2.9529 0.5618
20 0.3899 0.2212 1.1745 0.9419 1.2469
21 0 0 1.7281 21.1049 0.0819
22 0.0806 ? 2.1865 2.5177 0.8684
23 0.0001 0.0049 2.4603 8.8921 0.2767
24 0.9892 0.3072 0.0462 0.0108 4.2758
25 0.1074 0.1724 0.3406 2.2311 0.1527
26 0.664 0.2513 0.1573 0.4095 0.3841
27 0.8655 0.2215 0.1472 0.1444 1.0196
28 0.3899 ? 0.4533 0.9419 0.4812
29 0.9781 0.2495 0.0561 0.0222 2.531
30 0.769 0.2556 1.0257 0.2627 3.9045
31 0.6291 0.3509 1.7991 0.4635 3.8817
32 0.8233 0.2598 1.0635 0.1944 5.4705
33 0.4739 0.26 1.6474 0.7468 2.2058
34 0.5104 0.3191 0.3886 0.6725 0.5779
35 0.2173 0.183 0.485 1.5267 0.3177
36 0.7979 0.2642 1.0764 0.2258 4.7675
37 0.7 0.16 0.2598 0.3567 0.7282
38 0.0094 0.2267 0.8668 4.6642 0.1858
39 0.0806 0.2045 0.8122 2.5177 0.3226
Last Coefficient Update
0 -5.835E-8
1 1.401E-9
2 -8.597E-9
3 -2.822E-7
4 -4.566E-8
5 1.256E-7
6 1.058E-8
Covariate Means
0 5.65
1 56.575
2 15.65
3 0.35
4 0.275
5 0.125
6 0.525
Distinct Values For Each Class Variable
Variable 0: 1 2 3 4
Variable 1: 0 1
Stratum Numbers For Each Observation
0 1
1 1
2 1
3 1
4 1
5 1
6 1
7 1
8 1
9 1
10 1
11 1
12 1
13 1
14 1
15 1
16 1
17 1
18 1
19 1
20 1
21 1
22 1
23 1
24 1
25 1
26 1
27 1
28 1
29 1
30 1
31 1
32 1
33 1
34 1
35 1
36 1
37 1
38 1
39 1
Number of Missing Values = 0
Final Coefficient Matrix
0 1 2 3
0 -0.585 0.137 -4.272 0
1 -0.013 0.021 -0.634 0.526
2 0.001 0.012 0.064 0.949
3 -0.367 0.485 -0.757 0.449
4 -0.008 0.507 -0.015 0.988
5 1.113 0.633 1.758 0.079
6 0.38 0.406 0.936 0.349
Link to Java source.