A system and method for providing supervised learning to associate profiles in video audiences is provided. The method includes: receiving data providing an association of consumer profiles and video decoders to households within a network; recording zapping events (patterns) created by consumers; and associating zapping patterns of consumers with households. The step of associating further includes: collecting external data and converting a format of the external data into an internal format; converting zapping logs into different data models that can be used to provide set top box signatures; providing the set top box signatures; using the set top box signatures with a list of set top boxes and profiles to provide an association rule; and applying the association rule to the set top box signatures to determine a list of profiles of the consumer profiles associated with a specific set top box of the set top boxes.