To represent the voice samples in a parametric way, we have considered the cepstral representation as this has been found to be a more robust and reliable feature set for voice recognition than other forms of representation [16, 20].
We found that normality cannot be rejected (P < 0.05) for the majority of our samples; therefore, a parametric statistical analysis was selected.
In this analysis, due to the limited sample size, a parametric model, the Weibull distribution, is used for the underlying distribution of survival times.
The peak linkage signal from the entire sample was a parametric heterogeneity LOD score of 2.76 at rs1561817 (chr8: 26,591,503; hg18), located ~4 Mb telomeric of NRG1 and only ~20 kb centromeric of DPYSL2.
For comparison of paired samples, a non-parametric equivalent Wilcoxon signed-rank test was used.
In order to effectively fuse these different types of measurements, we proposed a Multi-scale Multi-rate Adaptive Sampling approach with a parametric description of the field[6].
In addition, for each miRNA-encoding DNA locus identified as altered in any of the samples, a separate non-parametric Wilcoxon rank sum test was applied to measure differences in expression between samples with deletions and amplifications, compared to samples with no CNAs.
A statistical model can also lead to physically unrealistic results whereas a dynamical model provides physically plausible limits, but a statistical model may reduce sampling uncertainties by providing a parametric description.
Due to the small sample size, we used parametric and a-parametric correlations (Table 2).
The small size of the samples precluded the use of a parametric t test.
A parametric independent samples t-test was used to compare the number of genuine oestrus events potentially observed between nulliparous females whose first oestrus coincided with a false oestrus, and nulliparous females from the same families who experienced their first oestrus event alone.
