Context Knowledge Extraction using Network Traffic Information

Abstract

A growing trend in information technology is not just reacting to changes, but anticipating them as much as possible. This paradigm is the base of modern applications, such as recommendation systems, the context-aware applications, among others. Anticipatory systems extend the idea to the communication systems, by studying patterns and periodicity in human behavior and network dynamic, to optimize the network performance. Particularly, for context-awareness applications is very important to extract autonomously contextual information. This work proposes a set of autonomic cycles of data analysis tasks to provide context awareness using the network traffic data, which gives information about the behavior of the traffic flow in a given context. This information about the network (links, users, type of traffic, etc.) is used to extract useful knowledge about the context using data analysis tasks.