If Hamming metric is adopted, the error will spread.
Namely, the localization performance can be improved using the second-step solution since it adopts the error distribution of the first-step estimate.
Somatic cells mostly adopt the error prone non-homologous end joining (NHEJ) pathway to repair DNA damage [39].
We first adopted the mean error rate (MER) and the mean-weighted error rate (MWER) [19], for the evaluation.
Adopting the trial-and-error constructive process described previously, we eventually trained and evaluated 514 MLPs: 30 of them with one hidden layer, with their layer's size ranging from 10 to 40 neurons, and 484 with two hidden layers, with hidden layers' sizes ranging from 5 to 27 neurons.
To demonstrate the effectiveness of this method, we will adopt the same error definition as [20].
For comparison, the EKF algorithm adopts the measurement error distribution as Gaussian distribution N 1,4).
For non-coherent networks, if the subspace distance is adopted, the propagated error will not be polluted 100% of the received packets which is better than Hamming distance.
We adopt the solution error measure ((f(x) - f(x*))), where f(x) is the best solution obtained by algorithms in one run and (f(x*)) is well-known global optimum of each benchmark function, which is recorded after 1.2e05, 6.0e05, and 3.0e06 function evaluations (FEs) for CEC'2010 and 5.0e03*D function evaluations (FEs) for CEC'2008, respectively.
Adopting the Multivariate Mean Square Error (MMSE) as optimization criteria, a robust solution could be found.
This risk, largely due to multiple tests (one at each locus), can be reduced by using a conservative significance level [ 10, 35], as done here by adopting the smallest type I error of our analysis which is the one calculated for logistic regressions α = 0.0006 (see later the justification).
