Figure 1.
Construction of Feature-dissimilarity Graph.
From the data matrix first Relevance Vector and Dissimilarity Matrix are Computed, then a weighted complete Feature-dissimilarity Graph is computed. Here an example of 5 feature-dissimilarity graph is depicted.
Table 1.
Algorithm 1: Graph based MObPSO (Minimization Problem).
Table 2.
Performance Analysis for Three Real-life Data Set.
Table 3.
10-fold Cross-validation Result Analysis for Three Real-life Data Set.
Figure 2.
The Heatmap of the gene markers for Prostate Cancer data.
The Heatmap describe the expression levels of the four up-regulated and two down-regulated gene markers for normal and cancerous type in Prostate Cancer data.
Figure 3.
The Heatmap of the gene markers for DLBCL data.
The Heatmap describe the expression levels of the three down-regulated gene markers for DLBCL and FL type in DLBCL data.
Figure 4.
The Heatmap of the gene markers for Child-ALL data.
The Heatmap describe the expression levels of the five down-regulated gene markers for after and before therapy in Child-ALL data.
Table 4.
Gene Markers Identified by the Proposed Method for Various Dataset.