Table 2 Performance of the DAMACY models on each dataset in the obese versus lean study, and the fusion of all three datasets.

From: A novel data fusion method for the effective analysis of multiple panels of flow cytometry data

Dataset

Accuracy

Sensitivity

Specificity

p-value

%increase

B cells

71%

80%

59%

<12/1000

1.5%

T cells

79%

86%

70%

<1/1000

4.5%

Innate

75%

75%

75%

<9/1000

11.8%

Data fusion

81%

82%

80%

<2/1000

 
  1. Accuracy indicates the percentage of correctly classified samples in the cross-validation study, sensitivity reflects accuracy in identifying the obese samples and specificity indicates the ability to detect control (lean) samples. The p-value is the relative amount of higher prediction accuracies found after 1000 permutations. The %increase values reflect how much the accuracy would increase if that dataset was included in the data fusion model when compared with a fusion model comprising only the two other datasets.
Back to article page