Skip to main content

Table 2 SVM classification

From: Transcriptomic signatures in whole blood of patients who acquire a chronic inflammatory response syndrome (CIRS) following an exposure to the marine toxin ciguatoxin

Identifier Gender Trained Predicted Confidence
Control 1-1 Female Control Control 0.754
Control 1-2 Female Control Control 0.764
Control 2-1 Male Control Control 0.796
Control 2-2 Male Control Control 0.789
Control 3-1 Male Control Control 0.794
Control 3-2 Male Control Control 0.829
Control 4-1 Male Control Control 0.805
Control 4-2 Male Control Control 0.832
Control 5-1 Male Control Control 1.000
Control 5-2 Male Control Control 1.000
Control 6-1 Male Control Control 0.826
Control 6-2 Male Control Control 0.821
Control 7-1 Female Control Control 0.877
Control 7-2 Female Control Control 0.827
Control 8-1 Male Control Control 0.710
Control 8-2 Male Control Control 0.772
Control 9-1 Male Control Control 0.796
Control 9-2 Male Control Control 0.936
Patient 1-1 Male Patient Patient 0.819
Patient 1-2 Male Patient Patient 0.805
Patient 2-1 Male Patient Patient 0.858
Patient 2-2 Male Patient Patient 0.847
Patient 3-1 Male Patient Patient 0.858
Patient 3-2 Male Patient Patient 0.917
Patient 4-1 Female Patient Patient 0.858
Patient 4-2 Female Patient Patient 0.853
Patient 5-1 Male Patient Patient 0.844
Patient 5-2 Male Patient Patient 0.813
Patient 6-1 Male Patient Patient 0.929
Patient 6-2 Male Patient Patient 0.898
Patient 7-1 Male Patient Patient 0.763
Patient 7-2 Male Patient Patient 0.852
Patient 8-1 Male Patient Patient 0.790
Patient 8-2 Male Patient Patient 0.821
Patient 9-1 Female Patient Patient 0.847
Patient 9-2 Female Patient Patient 0.904
*Control 10-1 Male   Control 0.449
*Control 10-2 Male   Control 0.399
*Control 11-1 Male   Control 0.627
*Control 11-2 Male   Control 0.660
*Patient 10-1 Male   Patient 0.019
*Patient 10-2 Male   Patient 0.108
*Patient 11-1 Male   Patient 0.976
*Patient 11-2 Male   Patient 1.000
  1. Results of predictions and confidence for cases of CIRS-ciguatera and controls using a Support Vector Machines classification algorithm, while withholding eight random data profiles* from training.