Fig. 3From: Novel feature selection method via kernel tensor decomposition for improved multi-omics data analysisBoxplot of \(u_{2j}\) when HOSVD is applied to a linear kernel computed using kidney cancer data. Left: TCGA, \(P=8.49 \times 10^{-47}\), right: GEO, \(P=4.07 \times 10^{-17}\). P values are based on the t test applied to \(u_{2j}\). T: tumors, N: normal kidney samplesBack to article page