Customer behavior model graph is a state transition graph that has one node for each possible service and transitions between these services; a probability is assigned to a transition between two services representing the frequency at which the user requested the services consecutively in the session.
Published in Chapter:
Broadband User Behavior Characterization
Humberto T. Marques Neto (Federal University of Minas Gerais (UFMG), Brazil), Leonardo C.D. Rocha (Federal University of Minas Gerais (UFMG), Brazil), Pedro H.C. Guerra (Federal University of Minas Gerais (UFMG), Brazil), Jussara M. Almeida (Federal University of Minas Gerais (UFMG), Brazil), Wagner Meira Jr. (Federal University of Minas Gerais (UFMG), Brazil), and Virgilio A.F. Almeida (Federal University of Minas Gerais (UFMG), Brazil)
Copyright: © 2008
|Pages: 19
DOI: 10.4018/978-1-59904-851-2.ch026
Abstract
This chapter presents a broadband user behavior characterization from an Internet service provider standpoint. Understanding these user behavior patterns is important to the development of more efficient applications for broadband users. Our characterization divides the users into two categories: residential and small-office/home-office (SOHO). It employs four characterization criteria: session arrival process, session duration, number of bytes transferred within a session, and user request patterns. Our results show that both residential and SOHO session interarrival times are exponentially distributed, and point out that a typical SOHO user session is longer and transfers a larger volume of data. Our analysis also uncovers two main groups of session request patterns within each user category: (i) sessions that comprise traditional Internet services, such as WWW services, e-mail, and instant messenger, and (ii) sessions that comprise peer-to-peer file sharing applications, basically. We also analyzed and classified the e-business services most commonly accessed by users, which did not vary significantly across the user categories.