Feature Selection for Video Traffic Identification with Competitive Swarm Optimizer

To identify video types in the network traffic, we identified the game video from a large number of videos as an imbalanced problem. An imbalanced data gravitation-based classifier (IDGC) was used to solve the classification of the video type in this study. We collected different types of videos and extracted the original feature dataset of two types of video traffic. We extracted 256 features and then used a competitive swarm optimizer (CSO) to select the effective feature from the origin feature data set. We selected the features by their weights. We experimented collected dataset. The developed feature selection method in this study successfully classified video traffic.