Serum and Cerebrospinal Fluid Cytokine Biomarkers for Diagnosis of Multiple Sclerosis

Figure 6

MS prediction accuracy using the KNN model with five randomly selected cytokines commonly expressed in patient serum as input. Serum common cytokines dataset consists of ten cytokines (IL-1α, IL-4, IL-18, CCL7, CCL27, INFγ, LIF, M-CSF, SCF, and TNFα). A: (IL-4, IL-18, CCL27, M-CSF, and SCF); B: (IL-4, CCL7, CCL27, INFγ, and SCF); C: (IL-1α, IL-18, INFγ, M-CSF, and TNFα); D: (IL-1α, IL-4, CCL27, INF-γ, and M-CSF); and E (IL-1α, IL-4, IL-18, LIF, and M-CSF) are datasets which are generated by considering five random cytokines. The accuracy is the average of 15 runs.