A Benchmark for Methods in Reverse Engineering and Model Discrimination: Problem Formulation and Solutions

Table 7.

Summary of Results of Nonlinear Model Analysis

Optimization Method
Approach
Case
State Variables
Stochastic
Gradient-Based
No weighting 1 M1 3.7195 × 106 3.7342 × 106
2 M2 9.0224 × 10−5 7.8716 × 10−5
3 E 3.6658 × 10−4 8.015 × 10−7
4 M1, M2 3.7198 × 106 3.7342 × 106
W, unity matrix 5 M1, E 3.7198 × 106 3.7342 × 106
6 M2, E 3.6658 × 10−4 7.9476 × 10−5
7 M1, M2, E 3.7195 × 106 3.7367 × 106
Weighted by square of average 8 M1 2.8287 2.8378
9 M2 0.0222 0.1201
10 E 0.0476 0.0493
W as in equation 24 11 M1, M2 2.8394 2.8428
12 M1, E 2.8685 2.8806
13 M2, E 0.0686 0.0692
14 M1, M2, E 2.8739 2.8857
Simplified Chen and Asprey 15 M1 1.356 3.6054
16 M2 36.3359 2.1768
17 E 2.6896 0.099
18 M1, M2 36.4986 7.6538
W as in equation 26 19 M1, E 2.6897 2.5091
20 M2, E 36.9019 2.4782
21 M1, M2, E 37.0645 7.763
Simplified Chen and Asprey without measurement variance 22 M1 24.88 22.9863
23 M2 1.5598 × 1011 1.4355 × 1011
24 E 3.6623 2.6398
25 M1, M2 1.5598 × 1011 1.4355 × 1011
W as in equation 29 26 M1, E 24.88 3.6468
27 M2, E 1.5598 × 1011 3.6207

28
M1, M2, E
1.5598 × 1011
1.4355 × 1011

This Article

  1. Genome Res. 14: 1773-1785

Preprint Server