ABSTRACT

Faceted browsing is widely used in Web e-Commerce sites. In these cases, an ordered list of
facets is often considered. This approach suffers from two main issues. First, one needs to spend
a significant amount of time to devise an effective list. Second, with a fixed list of facets it can
happen that a facet becomes waste if all products that match the query are associated to that
particular facet. Multifaceted search is a commonly used interaction paradigm in e-commerce
applications, such as Web shops. Because of the large amount of possible products, Web shops
usually make use of static information to determine which facets should be displayed.
Unfortunately, this approach does not consider the user query, leading to a non-optimal facet
drill down process. In this paper, we focus on automatic facet selection, with the aim of
minimizing the number of steps needed to find the desired product. We proposed several
algorithms for facet selection, which we evaluated against the state-of-the-art algorithms from
the literature. In this paper, our system is based on five components.

The first component of this paper is exploratory search in a faceted search interface.
Modern populations rely heavily on the world-wide web in searching information because it is
the largest human repository of knowledge. However, finding relevant information on the web is
often challenging task. In this work, we reviewed analyses and optimized the performances of
exploratory and faceted search techniques. A large amount of uncertainty about the goals of the
search is common in exploratory search. This study examined how searchers interacted with a
web- based, faceted library catalog when conducting exploratory searches. It includes eye
tracking, stimulated recall interviews, and direct observation to investigate important aspects of
gaze behavior in a faceted interface: what components of the interface searchers looked at, for
how long, and in what order. A limitation of these methods is that they
do not provide direct evidence of what elements of an interface searchers look at. This study
begins to fill that gap by the use of eye tracking analysis.

The second component in this paper is a novel collaborative-filtering-based recommender

system by user interest expansion via personalized ranking, named i-Expand. The goal is to build
an item-oriented model-based collaborative filtering framework. The i-Expand method
introduces a three layer, user–interests–item, representation scheme, which leads to more
accurate ranking recommendation results with less computation cost and helps the understanding
of the interactions among users, items, and user interests. Moreover, i-Expand deals with many
issues that exist in traditional collaborative filtering approaches, such as the overspecialization
problem and the cold-start problem. Finally, we evaluate i-Expand on three benchmark data sets,
and experimental results show that i-Expand can lead to better ranking performance than state-
of-the-art methods with a significant margin. The main limitation of i-Expand lies in its “bag of
items” assumption, where in each user’s rating record, the rating contextual information (e.g.,
rating time) is totally ignored.

The third component in this paper is the personalized facet search. This section proposes
an alternative evaluation methodology for faceted search based on user simulation and explicit
relevance feedback collected over a long period of time from many actual users. This
methodology is designed around the concept of measuring the utility of an interface based on
which a user performs in a search session. This new methodology can be used to evaluate
personalized faceted search interfaces in a quick, cheap and repeatable manner. This paper also
proposed a cheap evaluation methodology for personalized faceted search research. The
experimental environment is repeatable and controllable, which makes it a markable evaluation
environment. Although the simulated users differ from real users, the evaluation methodology
does provide insight into understanding how various faceted interface design algorithms perform.
This paper does not intend to claim whether this evaluation method is better or worse than user
studies. Instead, the outlined approach serves to complement user studies by being cheap,
repeatable, and controllable.

The fourth component in this paper is optimal facet selection which addressed the
problem of selecting the set of facets to be displayed alongside search results from a utility
perspective, aiming to minimize the effort of users in finding their target documents. We
modeled the manner by which users interact with faceted search engines, and proposed efficient
algorithms that select approximately optimal facets under those models, as selecting the optimal
set is NP-Hard. Our current selection algorithms do not explicitly consider the hierarchical
relationships among facets. For example, the algorithm might output both a facet and its
ancestor, despite its preference to select non-correlated facets.
 The fifth component in this paper proposes a method for ontology-based semantic search
systems. It specifically suggests a measure for dynamic selection of categories and an algorithm
to arrange them in an appropriate order. Finally, it proves the validity of the proposed approach
by using some evaluative measures. The objective of this paper is to propose a method for
generating dynamically categories refining search results onto ontology-based search systems. It
is thereby assumed that items resulted in search are structured data with their associated
properties. Another assumption made in this paper is that search results contain only one type of
exact item since ontology-based search systems aim at finding items that represent the user’s
intention and such items belong to an identical type.