Information Technology Reference
In-Depth Information
UBO has the clear goal to serve as a general model for interaction with an appli-
cation, a semantic form of the server log files. It is not intended to be part of a general
model for all possible types of behavior. The field theory of Lewin [ 20 ] proposes that
human behavior is the function of both the person and the environment. With UBO,
the focus is set on the environment, what type of application the user is interacting
with, what elements are visible to the user, etc. The user itself, their current emotions,
and needs are not part of UBO. This must be incorporated by other models.
As outlined in the previous section, the main goal of user behavior collection, the
web usage mining process, is to get detailed information about how users interact with
an application to understand what people want in order to offer better personalization
or recommendation. This has to be part of the UBO, too. The challenge is to build
a model that allows to manage explicit information such as a click event, and to
manage the implicit information that is also tracked with our tracking system.
7.3.1 Conception of UBO
UBO orients itself on standard log file formats. As stated above, its purpose is to
provide a semantic model for user behavior that can be extended with additional
meta-information. Existing work on general user behavior ontologies is scarce. The
work of Schmidt et al. [ 35 ] proposes a set of different models to capture all relevant
data for website personalization. The used models cover the website structure (Web
Portal Ontology), website content (Content Ontology), user profiles (User Ontol-
ogy), and website usage data as well as knowledge about the adaptation process
itself (Adaptation Ontology). The most important ontology is the adaption ontology,
which is used to decide if an adaptation should take place and how to do that. The
adaption decisions are based on predefined rules. The ontology most related to the
UBO is the Behavior Ontology [ 35 ]. This ontology captures atomic events, such
as mouse related or keyboard events, and when an event started and ended. UBO
centers around the Element a user is interacting with. With UBO, the interaction
with that of a website element is tracked, how the user interacted with the element,
and also what other elements were visible and semantically connected. UBO allows
collecting more information than the pure Behavior Ontology presented in [ 35 ]. The
combination of the Web Portal and the Behavior Ontology allows at least connecting
an event to the page structure, but still the possibility to track underlying semantic
connections on a webpage is not given. It is also not explained how the Web Portal
Ontology copes with partial reloads of the website. This change in the website struc-
ture is trackable with our semantic tracking solution and can be captured using UBO.
Ngoc et al. [ 24 ] present an approach for generalized ontologies for user preferences,
the Spatio-Temporal Ontology of User Preference (STOUP), and behavior routine,
the Spatio-Temporal Ontology of User Routine (STOUR). STOUR covers part of the
intended UBO functionality as it allows to model reoccurring activities in a Routine
Element connected with time and system information. This is a higher aggregation
of the UBO Event Element but already processed to meta-knowledge. The goal of
Search WWH ::




Custom Search