A-2006-2 Janne Jämsen, Timo Niemi and Kalervo Järvelin,
Using derived typesin discovering and analysing semantic associations. September 2006.
Abstract.
Semantic associations are direct or indirect linkages between two entities that
are con-strued from existing associations among entities. In this paper we
extend our previous query lan-guage approach for discovering semantic
associations with an ability to retrieve semantic associa-tions that, besides
explicitly stated (base) associations, may contain associations derived using
logic-based derivation rules. As will be shown, this makes it possible to find
semantic associations that are both compact and intuitive. To implement this new
feature, we introduce a rewriting princi-ple that utilizes derived associations
to reduce resulting semantic associations if possible. Other proposed means to
assist the interpretation of query results include answer expansion and the
order-ing of answers. The incorporated answer expansion feature lets the user
investigate rewritten se-mantic associations in a query result at the desired
level of detail. The ordering of answers is based on the lengths of the
resulting semantic associations, whereby priority is given to shorter semantic
associations which often express close and relevant relationships.
Keywords: semantic association, query language, logic, derivation rule, deductive
database, answer expansion.
A-2006-3 Jaakko Hakulinen, Markku Turunen and Kari-Jouko Räihä,
Tutoring in a Spoken Language Dialogue System. September 2006.
Abstract.
We have developed interactive software tutors to teach users how to use a
spoken dialogue timetable system. The tutors teach the
functionality and interaction style of the telephone-based timetable system
to new users by guiding users and monitoring their
interaction. The primary modality of the tutors is graphics and they
feature a visual representation of the spoken dialogue between
a user and the system. Two different versions of tutoring were compared to
a static web manual with the same information in a
between-subjects experiment with 27 participants. Participants’ evaluations
of guidance materials were the most positive towards a
tutor featuring a graphical interface representation of the timetable
query. An otherwise similar tutor, which did not have the
graphical user interface representation, received the weakest evaluations.
Error rate variances suggest that tutoring is better than
static guidance especially for those who most need guidance.
pdf-file
A-2006-4 Turkka Näppilä, Kalervo Järvelin and Timo Niemi,
Construction of data cubes from structurally heterogeneous XML document
collections. October 2006.
Abstract.
Advanced data analysis methods like OLAP, Knowledge Discovery, and Data Mining
try to meet the information needs of modern organizations. OLAP (Online
Analytical Processing) is a powerful tool for analyzing multidimensional data
cubes, often on an ad hoc basis. As organizations today are increasingly
connected through the Web, the data for OLAP data cubes must often be integrated
from distributed and autonomous information sources. In this case, severe
problems occur due to semantic, syntactic, and structural data heterogeneity.
While XML provides a standard data exchange format, the problems of
heterogeneity remain. Popular path-oriented XML query languages, such as XQuery,
require the user to know, in much detail, the structure of the documents to be
processed. This paper demonstrates that such XML query languages are laborious
and troublesome in data integration. It is also argued that the proposed
improvements, Lowest Common Ancestor (LCA) -based query evaluation strategies,
are insufficient. The paper introduces both a novel high-level data extraction
primitive utilizing the developed Smallest Possible Context (SPC) evaluation
strategy and an advanced OLAP data cube construction operation. The paper
demonstrates, through a system prototype and a sample application in
Informetrics, that the approach is a real improvement in data integration.
A-2006-5 Marko Junkkari, Paavo Arvola and Jaana Kekäläinen,
Grammatical Approach to XML Information Retrieval Query Languages. November 2006.
Abstract.
The formal syntax and semantics of a structured query
language, XIL, for document-oriented XML retrieval are given. The
syntax is represented by a context free grammar and the semantics by
an attribute grammar. Dewey like structural indices and traditional
set theoretical operations for their manipulation are utilized as a
meta-language for the attribute grammar. These indices allow effective
construction of relevance-based views for result representation.
Although we follow the syntax of SQL our aim is not to present an
extension to any existing query language. Instead, we show how an SQL
like query language can be applied to XML retrieval without any
additional constructors - even such that have been established in SQL
for flat structures. Further, the language does not involve difficult
constructs, like shared variables but still has the restructuring
power typical for SQL. Structure and structural conditions are given
with path expressions. Further, simple content-only queries, as well
as querying in heterogeneous structures are supported.