Figure 3
From: Precise diagnosis of three top cancers using dbGaP data
Classification performance of NBCs over the six datasets from five dbGaP studies. (A) Accuracy, sensitivity and specification of \({\texttt {NBC}_{\texttt {634}}^{\texttt {(1)}}}\), \({\texttt {NBC}_{\texttt {753}}^{\texttt {(1)}}}\), \({\texttt {NBC}_{\texttt {147}}^{\texttt {(1)}}}\), \({\texttt {NBC}_{\texttt {517}}^{\texttt {(1)}}}\), \({\texttt {NBC}_{\texttt {306}\hbox {-}{} \texttt {JL}}^{\texttt {(1)}}}\) and \({\texttt {NBC}_{\texttt {306}\hbox {-}{} \texttt {AA}}^{\texttt {(1)}}}\) according to 10 fold cross-validation, where the error bars are in form of “mean±std” computed by repeatedly performing 10-fold cross-validation for 100 times. (B) Performance of \({\texttt {NBC}_{\texttt {634}}^{\texttt {(2)}}}\), \({\texttt {NBC}_{\texttt {753}}^{\texttt {(2)}}}\), \({\texttt {NBC}_{\texttt {147}}^{\texttt {(2)}}}\), \({\texttt {NBC}_{\texttt {517}}^{\texttt {(2)}}}\), \({\texttt {NBC}_{\texttt {306}\hbox {-}{} \texttt {JL}}^{\texttt {(2)}}}\) and \({\texttt {NBC}_{\texttt {306}\hbox {-}{} \texttt {AA}}^{\texttt {(2)}}}\) according to 10-fold cross-validation. (C) Accuracy (accu.; %), sensitivity (sens.; %) and specification (spec.; %) of each \({\texttt {NBC}_{\texttt {study}}^{\texttt {(1)}}}\) according to leave-one-out. (D) Performance of each \({\texttt {NBC}_{\texttt {study}}^{\texttt {(2)}}}\) according to leave-one-out.
