Biomedical Engineering Reference
In-Depth Information
Table 2. BIC, AIC and Loglikelihood Values for 2, 3 and 4 Stages Models
Model
BIC
AIC
Log of Likelihood
P-value for
Values = log L(
^
jY )
Goodness of t
Homogeneous 2-stage
188.93
182.07
-87.04
0.10
None-Homogeneous 2-stage
177.48
167.20
-77.60
0.42
Homogeneous 3-stage
151.42
139.42
-62.71
0.98
Homogeneous 4-stage
150.76
133.61
-56.81
0.99
Homogeneous 5-stage
211.67
189.38
-81.69
0.06
P
P
n
j=1
fy
ij
^
ij
g
2
=
^
ij
2
(72k),
(k = the number of parameters under the model)
(2) log L(
^
jY ) =
m
i=1
Notes: (1)The p-value is based on
P
P
m
i=1
n
j=1
f
^
ij
+ y
ij
log
^
ij
log (y
ij
!)g;
^
ij
= n
ij
p
i
(j)
Table 3. Estimates of Parameters for the 4 Stage Homogeneous Model
with Predictive
CIGARETTES/DAY(RANGE AND MEAN)
0
1-4
5-9
10-14
15-19
Parameters
0
2.7
6.6
11.3
16
0
(i)
218:03 247:95 286:41 285:07 295:57
12:81 15:79 16:79 16:04 16:73
b
1
(i)
5:93E02
6:72E02
7:17E02
7:47E02
7:66E02
8:77E05 5:40E05 3:43E05 1:46E05
1:16E05
d
1
(i)
6:18E02
6:02E02
5:99E02
5:95E02
5:93E02
9:68E05 4:78E05 2:90E05 1:06E05
1:50E05
b
1
(i)d
1
(i)1:31E03
1:78E03
3:95E03
5:21E03
6:21E03
5:60E05 6:07E05 6:64E05 4:12E05
3:17E05
1
(i) 5:29E05 6:74E05 5:48E05 5:68E05 5:52E05
2:67E06 1:67E06 1:12E06 3:96E07 3:65E07
b
2
(i) 8:08E02 8:92E02 8:97E02 9:14E02 9:34E02
1:59E03 7:34E04 5:04E04 2:10E04 2:62E04
d
2
(i) 8:84E02 8:63E02 9:00E02 9:19E02 9:37E02
1:99E03 6:82E04 5:52E04 2:15E04 2:15E04
b
2
(i)d
2
(i)7:60E03 2:90E03 3:00E04 5:00E04 3:00E04
2:55E03 1:00E03 7:47E03 3:01E04
3:39E04
2
(i)
4:01E05 5:55E06 3:28E06 3:84E05 1:61E05
3:86E05 4:99E06 3:70E06 4:75E06 3:37E06
b
3
(i)
1:08E02
9:32E02
4:79E02
9:84E02
1:01E01
1:07E02 2:53E04 1:65E02 7:63E04
4:89E04
d
3
(i)
7:18E02
6:27E02
1:10E01
8:59E02
8:16E02
2:45E02 1:68E04 2:45E02 5:60E04
4:49E04
b
3
(i)d
3
(i)2:21E03
7:88E03
1:46E02
1:97E02
2:11E02
2:42E03 2:71E03 1:84E03 4:78E04
5:79E04
3
(i) =
3
9:54E062:51E08
Search WWH ::
Custom Search