A Two-Level Model for the Analysis of Syndrome of Acute Ischemic Stroke: From Diagnostic Model to Molecular Mechanism

Figure 2

The accuracies of classification schemes based on different groups of symptoms. Symptoms whose information gains are above the threshold are selected. The black line indicates accuracies of models based on different groups of symptoms, which helps to select core symptoms. The red line presents the accuracy of the original classification scheme without feature selection as a comparison.