The goodness-of-fit results indicated that the best fitting distribution is different for individual homogeneous regions.
It was found for all cases that the best fitting corresponds to lognormal distributions.
By using the least squares technique, it was observed that the best fitting occurs for m = 1/2.
Confirmatory factorial analyses suggest that the best fitting model includes four specific relational factors at the first level and two factors at the second level, these last two factors representing parental and peer attachments.
Unlike the HKM and FCM described above, where the number of clusters is equal to the number of sources, the GMM-based clustering methods have the additional complexity in that the best fitting for the data set to a mixture model may not necessitate that P is equivalent to the number of sources [14].
We previously reported, however, that the best fitting model for either log(EXT) or log(TOF) contained three separate growth rates, separated by two change points [14].
In fact, from Eq. 2, we found that, the best fitting values of the relaxation time τ, which were extracted from simulated time courses of extension (Fig. 3 A and B), were about 1.5∼3.2 ps, corresponding to the relative extension x/L of about 1% for HP(2 20) and its four analogues.
Fractional polynomials indicated that the best fitting final model was achieved with linear values.
Application of Akaike information criterion across 12 possible models indicated that the best fitting model for cyt b was GTR + G + I and for the two introns, HKY.
We found that the best fitting model, as indicated by the smallest deviance information criterion, was the Bernoulli-Gamma model including φ j.
Using mean plots and one way analysis of variance tests, we observed that the best fitting trend lines for these associations were linear.
