*2.6. SVM Discriminant Analysis to Predict Patients with Seasonal Influenza based on the Three Vital Signs Measured*

Aiming at screening using features of HR, RR and body temperature of patients with infection, we proposed a classification model based on SVM. SVM is a method that predicts the separating hyperplane to maximize the margin between the two classes and achieves a high generalization capability. The SVM discriminant function is defined as

$$\begin{array}{c} \min\_{w,\ \ w\_0,\ \xi} \left( \frac{1}{2} \|w\|^2 + C \sum\_{i=0}^{N} \xi\_i \right) \\ \text{subject to} \begin{cases} \ y\_i f(x\_i) \ge 1 - \xi\_i \\ \xi\_i \ge 0 \end{cases} \end{array} \tag{7}$$

where *w* is a constant that indicates the SVM coefficients corresponding to HR, RR and temperature; *yi* is a category of health or infection; *C* is the penalty parameter and ξ*<sup>i</sup>* is the slack parameter; *f*(*xi*) is linear discriminant function formula *w* · *xi* + *w*0. The calculation of SVM is performed using the MATLAB software.
