For the classification of samples, cluster analyses were performed on a Bray-Curtis dissimilarity distance matrix of standardized genera abundance data, derived from 16S rDNA amplicon libraries generated from extracted RNA (cDNA) or DNA, based on group average by the Unweighted Pair Groups Method using Arithmetic means agglomeration algorithm (UPGMA).
We analyzed expression of the candidate genes and used linear discriminant analysis for the classification of samples.
Similarly, in the Array6.0-only analysis, the average testing accuracies were only 0.70 and 0.79 for the classification of samples from "CHB and JPT" and "four populations", respectively.
A highly significant (p = 2.2 × 10−5) OPLS model could be built for the classification of samples by production year (model 8).
The average testing accuracies ranged from 0.53 to 0.79 for SNP-only analyses and increased to around 0.90 when GE markers were integrated together with SNP markers for the classification of samples from closely related Asian populations.
The majority of the classification analyses produced an average testing accuracy, calculated over 10 cross-validation datasets, greater than or close to 90%, with the exception of two SNP-only analyses; the 500 K-only and Array6.0-only analyses had relatively low testing accuracies for the classification of samples from two closely related ethnic populations, CHB and JPT.
However, if GE markers also were integrated together with SNP markers for the classification of samples from "CHB and JPT" and "four populations", the average testing accuracies increased to 0.89 and 0.92, respectively, in the 500 K + GE analysis and to 0.92 and 0.91 in the Array6.0 + GE analysis.
Multivariate methods were used for patterns recognition and the classification of samples.
For the classification of a sample x with the calibrated classifier each stage c t updates consecutively the sample trace.
Figure 3 Differences between a cascade and a McCascade for the classification of a sample x.
Experimental results have shown that the higher layer visualizing feature extraction and the transfer learning deep networks are effective for the classification of small sample target objects in the sky.
