Information Technology Reference
In-Depth Information
Although all of our test persons make use of standard search engines, most of them
can imagine to use our system at least in combination with a search engine on their own
mobile devices. The iPhone test group even would use our system as their main search
tool (on the smartphone) when the proposed improvements have been implemented.
7
Related Work
Our approach is unique in the sense that it combines interactive topic graph extraction
and exploration on different mobile devices with recently developed technology from
exploratory search, text mining and information extraction methods. As such, it learns
from and shares ideas with other research results. The most relevant ones are briefly
discussed below.
Exploratory Search.
[10] distinguishes three types of search activities: a) lookup search,
b) searching to learn, and c) investigative search, where b) and c) are considered as
forms of exploratory search activities. Lookup search corresponds to fact retrieval,
where the goal is to find precise results for carefully specified questions with mini-
mal need for examinating and validating the result set. The learn search activity can
be found in situations where the found material is used to develop new knowledge and
basically involves multiple iterations of search. It is assumed that the returned set of
objects maybe instantiated in various media, e.g., graphs, maps or texts. Investigative
searching is a next level of search activity that supports investigation into a specific
topic of interest. It also involves multiple iterations even for very long periods and the
results are usually strictly assessed before they are integrated into knowledge bases.
Our proposed approach of exploratory search belongs to the searching to learn activity.
In this spirit, our approach is more concerned with recall (maximizing the number of
possibly relevant associated topics that are determined) than precision (minimizing the
number of possibly irrelevant associated topics that are determined).
Collocation Extraction.
We consider the extraction of a topic graph as a specific
em-
pirical collocation extraction task
. However, instead of extracting collocations between
words, which is still the dominating approach in collocation extraction research (e.g.,
[2]), we are extracting collocations between chunks, i.e., word sequences. Furthermore,
our measure of association strength takes into account the distance between chunks and
combines it with the PMI (pointwise mutual information) approach [15].
[6] also exploit the benefit of Web snippets for improved internet search by grouping
the web snippets returned by auxiliary search engines into disjoint labeled clusters. As
we do, they also consider methods for automatic labeling. However, their focus is on im-
proving clustering of terms and not on the extraction of empirical collocations between
individual terms. Furthermore, they advocate the “document-comes-first” approach of
clustering Web snippets which is inappropriate for our methodology, cf. sec. 4.
Unsupervised Information Extraction.
Web-based approaches to unsupervised infor-
mation extraction have been developed by Oren Etzioni and colleagues, cf. [1]; [5];
[16]. They developed a range of systems (e.g., KnowItAll, Textrunner, Resolver) aimed
