<p>(A) Distribution of scores of the SVM method. The vertical line indicates the cutoff for the selection of features for the clonal model. The scores of the top-scoring features are listed. (B) Distribution of scores of the Lasso method. Top-scoring features in the distribution are indicated. On both panels, positions of the features mapped on the V3 loop structure are indicated in brackets, labels are colored according to the clusters shown in <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1002977#pcbi-1002977-g004" target="_blank">Figure 4</a>.</p
<p>(A) Models are categorized by the type of features they use. Boxes indicate the 25<sup>th</sup> (...
Ranked accuracies from the support vector machine analysis described in the paper for all of the cla...
<p>The x-axis show the parameter ranges comprised by each of the 10 bins. The y-axis shows the absol...
<p>ROCR of different models are plotted. The legend lists the number of features, AUC and sensitivit...
<div><p>(A) The gray shading indicates prediction accuracy as a function of SVM score (left <i>y</i>...
<p>Upper panel, frequency distributions of classification patterns identified by the SVM composite m...
<p>Histogram showing the predictive distributions (probability mass functions) on the number of fall...
Sparsity or parsimony of statistical models is crucial for their proper interpretations, as in scie...
<p>Ridge and lasso models were fit from the complete network. The resulting standardized coefficient...
<p>Rank and feature scores obtained after performing feature selection show the structural profile m...
<p>The figure shows the error rate of SVM vs. top ranked discriminant time-frequency features. Simil...
<p>Distribution of the top 1% connections contributing to correct SVM classification for (A) S vs. W...
<p>The comparison is done separately for each classifier: (A) SVM-linear (B) SVM-radial (C) Random F...
LDA feature selection (a) selects 192 features with most being relative PSD band features and standa...
<p>Note:*All heartbeats that matched the reference template are assigned to SVB-class by Stage1.</p>...
<p>(A) Models are categorized by the type of features they use. Boxes indicate the 25<sup>th</sup> (...
Ranked accuracies from the support vector machine analysis described in the paper for all of the cla...
<p>The x-axis show the parameter ranges comprised by each of the 10 bins. The y-axis shows the absol...
<p>ROCR of different models are plotted. The legend lists the number of features, AUC and sensitivit...
<div><p>(A) The gray shading indicates prediction accuracy as a function of SVM score (left <i>y</i>...
<p>Upper panel, frequency distributions of classification patterns identified by the SVM composite m...
<p>Histogram showing the predictive distributions (probability mass functions) on the number of fall...
Sparsity or parsimony of statistical models is crucial for their proper interpretations, as in scie...
<p>Ridge and lasso models were fit from the complete network. The resulting standardized coefficient...
<p>Rank and feature scores obtained after performing feature selection show the structural profile m...
<p>The figure shows the error rate of SVM vs. top ranked discriminant time-frequency features. Simil...
<p>Distribution of the top 1% connections contributing to correct SVM classification for (A) S vs. W...
<p>The comparison is done separately for each classifier: (A) SVM-linear (B) SVM-radial (C) Random F...
LDA feature selection (a) selects 192 features with most being relative PSD band features and standa...
<p>Note:*All heartbeats that matched the reference template are assigned to SVB-class by Stage1.</p>...
<p>(A) Models are categorized by the type of features they use. Boxes indicate the 25<sup>th</sup> (...
Ranked accuracies from the support vector machine analysis described in the paper for all of the cla...
<p>The x-axis show the parameter ranges comprised by each of the 10 bins. The y-axis shows the absol...