Safety risk control detection method based on improved SVM algorithm dynamics

The invention particularly relates to a dynamic safety risk control detection method based on an improved SVM (Support Vector Machine) algorithm. According to the dynamic security risk control detection method based on the improved SVM algorithm, the SVM algorithm is improved, so that a security risk control system can actively identify 0day attacks, and the detection rate of the 0day attacks is improved; when a certain attack vector is judged to be a determined unknown attack vector, an attack feature code is automatically extracted, a detection rule is formed, the extracted feature code can be updated to an existing feature library, artificial participation is reduced, linkage with other cloud centers is carried out in time, and the 0day attack is effectively prevented. According to the dynamic security risk control detection method based on the improved SVM algorithm, unknown attacks can be actively identified, the artificial participation cost is reduced, the overall detection rate of the security risk co