Information Technology Reference
In-Depth Information
Due to the huge dimensions of the random constants' domains, it is impor-
tant to be able to control the degree of genetic variation in them. Conse-
quently, special mutation operators were implemented that allow the autono-
mous control of the mutation rate both on the head/tail domain and on the Dc
domain. The inversion and IS transposition operators are also controlled sepa-
rately and, therefore, a Dc-specific inversion and a Dc-specific transposition
were also implemented in the GEP-KGP algorithm. The other operators (one-
point and two-point recombination, gene recombination, and gene transposi-
tion) remain unchanged and work exactly as described in chapter 5.
7.3 Evaluating the Performance of STROGANOFF
In this section, we are going to evaluate the performance of three STROGA-
NOFF systems (both the original and the enhanced STROGANOFF, and a
multigenic implementation of the enhanced STROGANOFF) by comparing
them with four simpler GEP systems (the basic GEA both with and without
RNCs and the cellular system both with and without RNCs) on the sunspots
problem. The performance of all the different systems will be compared in
terms of average best-of-run fitness and average best-of-run R-square.
For this prediction task, 100 observations of the Wolfer sunspots series
were used (Table 7.3) with an embedding dimension of 10 and a delay time
of one (see section 7.4 below to learn how to prepare time series data for
symbolic regression). Thus, for the sunspots series presented in Table 7.3,
Table 7.3
Wolfer sunspots series (read by rows).
101
82
66
35
31
7
20
92
154
125
85
68
38
23
10
24
83
132
131
118
90
67
60
47
41
21
16
6
4
7
14
34
45
43
48
42
28
10
8
2
0
1
5
12
14
35
46
41
30
24
16
7
4
2
8
17
36
50
62
67
71
48
28
8
13
57
122
138
103
86
63
37
24
11
15
40
62
98
124
96
66
64
54
39
21
7
4
23
55
94
96
77
59
44
47
30
16
7
37
74
