Algorithm 3: FHR Classsification using Support Vector Machine

Input: $\Delta =\left\{\left(d_1,y_1\right),\dots .,\left(d_n,y_n\right)\right\}$, class labels

$C=\left[c_1|,c_2,c_3\right]$, initial weight vector $w={\left[w_i\right]}^T$, hyper-parameters, training set $T$

Step1: Define the hyperplane $w^T\ast d+b\leftarrow 0$

Step3: Maximize the margin

$w\ast d+b\vee \frac{w}{\left|w\right|\vee \leftarrow \frac{1}{\left|w\right|\vee }}$

between the hyperplane and the plane median

Step4: Minimize $\left|w\right|\vee$ to maximize the margin

Output: $w_{T+1}$