nPool: Cross-Validation in EC-Star 87
Fig. 8 Distribution of
originating segments for 100
fittest candidates in top-layer
of runs with 16 segments.
Note that segments 1, 9, and
12 have no candidates
represented in the top 100
Hundred Best Genes Distribution Among
Thirty Two Segments
32(^31) 0%0%
30
8%
1
2% 12%^2
3
10%
4
2%
5
2%
7
(^10) 1%^8 2%
1%^13 0%^12 0%
14
1%
15
1%
6
14%
9
4%^11 2%
16
12%
17
0%
19
0%
20
0%
21
1%
22
3%
23
2%
24
2%
25
1%
26
0%
28
0%
27
4%
29
7%
18
6%
Fig. 9 Distribution of originating segments for 100 fittest candidates in top-layer of runs with 32
segments. Note that segments 10, 12, 17, 19, 20, 26, 28, 31 and 32 have no candidates represented
in the top 100
4.2 Discussion
The experiments show that running in nPool mode does not adversely impact the
quality of the results, and, there’s a hint that it actually may be helping to improve
them. This may be due to differences in the makeup of the originating segments,
allowing for better generalization. We base this on the evidence of disparity in the
contribution of different originating segments to the make-up of the top performers.