Voice and video calls constitute an increasingly important class of traffic that is carried over WiFi. We present a generic codec independent method for measuring the quality of voice and video calls (Skype, GotoMeeting, Hangouts etc.) using supervised machine learning.
The method involves building a regression model that provides call quality index using training data set comprising of parameters extracted from real data. The regression model is then applied on parameters extracted from live calls by the Access Points that are in the traffic path to measure the quality index of each call. The results aggregated and presented on the cloud dashboard provide real-time health of voice and video calls carried over the WiFi deployment.