to home
Vectorial Web Search
General similarity search of quantifiable resources is possible on the
Web, if these are represented in standardized way. As standardized data
structure for representation of quantifiable resources a "Vectorial Resource
Descriptor" or "VRD" is proposed, which contains a feature vector for representation
of the quantitative data and a "Vector Space Identifier" or "VSI" (a HTTP
URI) which uniquely identifies the meaning of every dimension (number)
of the feature vector. Feature vectors of VRDs with the same VSI are directly
comparable using a given metric. At this similarities of the resources'
data are mapped to spatial similarities of the feature vectors. So similarity
search is possible by calculating distances: resources are the more similar,
the smaller the distance between the feature vectors of the representing
VRDs is. This vectorial (numeric) similarity search could be efficiently
combined with conventional word based search: Realization of vectorial web search
(2009).
Talk (2010 presented
here)
Application in medicine (2010 published
here)
Additional information:
A
Searchable Patient Record Database for Decision Support (2009, also
here
accessible)
Design
of a global medical database which is searchable by human diagnostic patterns*
(2008, also here accessible)
Short video
(2008, with old nomenclature using the old word "pattern" which is now replaced by "VRD"*)
*Change of nomenclature and data representation since 10. October 2009,
please contact me in case of interest in details.
Many numeric data are stored in separated databases in the hidden web.
The proposed VRD structure gives also motivation to pack such numeric data
in a globally accessible and interchangeable form. So it could also help
to make hidden (numeric) data from the deep web accessible for all people.
Keywords:
VRD, Vectorial Resource Descriptor, VSI, Vector Space Identifier,
QRI, HTTP URI, feature vector, Vectorial Web Search