When the fitting was poor, fits were redone by adjusting the initial parameters.
The first starting point is given by the initial parameter vector.
Maximize T in (A.1) with respect to θ that is on the one-dimensional straight line determined by the initial parameter vector θ0 and the gradient vector u (Linear Search; e.g., Kowalik and Osborne, 1968).
Our tests show that the initial parameters produced by the weighted vote method cause almost 50% of child nodes to fail the conditional independence test, but as the EM algorithm learns more likely parameter settings, fewer and fewer nodes fail the test.
The number of parameters set npt is fixed as the maximum of the domain of variation of the initial parameters divided by their uncertainty u in the process of fabrication [ 38 ]: npt = max { [ max (r 1 0 ) − min (r 1 0 ) ] / u (r 1 ) ; [max(e ) − min(e )] /u (e ); [ max (n 1 0 ) − min (n 1 0 ) ] / u (n 1 ) ; [max − min] /u }.
Here, the initial parameters and are given by and, The sequences and are connected by the following formulas: (2.22).
The first step of the proposed method is to determine the number of hidden layer nodes and the initial parameters of the kernel by the SVR method.
(i), for all (ii) There exist two complex numbers and with such that (2.20) (2.21) Here, the initial parameters and are given by and, (iii) The sequences and are connected by the following formulas: (2.22) (2.23) ., for all.
The Markov chain was initialized by randomly selecting the initial parameter value and each Markov chain was sampled every 100 cycles for 1,000,000 generations with a burn in of 100,000 cycles.
Furthermore, the reference parameter set θ j 0 (equation 6) is exchanged in this case by θ j 1 (the initial parameter set representing phenotype B) in order to regularize the backward trajectory.
