Intelligent Information Retrieval in the Deep Web: an adaptable semantic model for retrieving, indexing and visualising Web Knowledge
- When?
- Thursday 9 December 2010, 11:00 to 12:00
- Where?
- 39BB02
- Open to:
- Staff, Students
- Speaker:
- Mr Georgios Michalakidis
Humans communicate through different signals. Due to an ability to perceive things, our species can perform this communication accurately and efficiently even when noise is introduced or the signal is presented in different formats. Computer technologies aid in cognitive processing and can, to some degree, support intellectual performance and enrich individuals’ minds. We operate (and design systems that operate) using analogies, such as reasoning, comparisons, and synonymity. In the end: is analogy a shared abstraction? Does it derive from mathematics? Is it high-level perception in shared structure theory?
This research uses analogies in mathematics and linguistics to help generate intelligent information extraction mechanisms and decision-support systems of measurable quality. The research exploits the unique characteristics of associations in data, through annotation and metadata generation, while addressing issues like the overload of information, stovepipe systems and poor current-generation content aggregation. For this purpose, and in an effort to generate the cognition rationale behind a confidence-assigning method, WebCandY has been developed; an algorithm for domain identification using fuzzy entity-name input lists.
Tests are being run for comparing RDF’s applicability in exchanging semantically annotated data (with OWL on the ontology layer and SPARQL for querying) against an RDB-Metadata approach (with time-series for evaluating change in the data and elastic lists for visualisation). Company profiling is being used as an exemplar due to the content availability and the change of data in time.
The preliminary in silico experimentation shows that this research can compete with RDF in terms of richness of the datasets and efficiency in both storing and retrieving data. More comparators will be introduced, and data evaluated and analysed statistically. The solution proposed is generating the metadata annotations and knowledge representation model automatically through identification and indexing of collateral data with an aim to contribute to and fuel further research into decision-support or, ideally, decision-making systems with the least required human intervention.
