Extraction and Classification of User Behavior

The multimedia document generator, iClass system, has been used by professors from the Institutes of Chemistry, Mathematics and Computer Science from the University of São Paulo aiming at helping the multimedia content production and availability. Data from user interactions, available in iClass system, have motivated this work which aims at studying the user behavior under different circumstances. The behavior extracted makes possible the analyze of different patterns of the same user, among groups, and distinct users. Those pattern differences should help to understand user evolution in iClass system under diverse situations. The data are grouped by a neural network and afterwards Markov Chains are built to represent their behaviors in different time moments. The detected user or group behavior variations are related to classify profiles and comprehend them in different situations.