Support Vector Machine-Based Classification of Malicious Users in Cognitive Radio Networks

Algorithm 1. Proposed SVM-based MU classfication algorithm.

 A. Generation of Data

 Initialization of parameters such as number of iteration, number of Sus.

 Generate random MUs with Gaussian distribution.

 Generate normal SUs.

 Generate the indices on which MUs attack.

 Generate indices for position of normal SUs.

 B. Sensing the data

  For to Sensing Interval

   For to N

    Energy reported by the SUs.

   End

   For to M

    Energy reported by the MUs.

   End

 End sensing interval

 C. Support Vector Machine Algorithm

    1. Data Processing

     i. Combining the data

     ii. Input the data

     iii. Train the data

     iv. Find the number of examples and attributes used in the data.

     v. Extract the attribute matrix X and the label vect Y.

    2. Support Vector

     i. Finding the support vectors (Corner points)

     ii. Draw the upper and lower hyperplanes.

     iii. Finding the maximal margin by the upper and lower hyperplane.

    3. Classification

     i. Linear

      Finding weights

     ii. Drawing the hyperplane to classify the data.

 D. Plotting

  i. Normal data plotting.

  ii. Malicious data plotting.

  iii. Plotting of hyperplane.

Algorithm 1. Proposed SVM-based MU classfication algorithm.