III Conference on Andalusian e-Science
Paper: Parsalyzer, a new tool for improving Internet searchesPonencia:
Author: Fernando Moreno-Torres Camy (MTC SOFT)
SummaryThe amount of information available on Internet is so huge that keyword searches return so many erroneous documents that it makes them ineffective, especially in cases which use common search terms. A solution to this problem is to “enrich” the texts before indexing them, so as the additional information can be used to elaborate more selective indexes which give search engines greater filtering possibilities.
By using an exhausting linguistic analysis we’ve developed a computer program, PARSALYSER, which effectively labels every word with its linguistic attributes. PARSALYZER achieves this labelling automatically and at a very low computational cost. The texts enriched with this software represent a much more valuable raw material than simple plain text without attributes.
They open up a whole set of new possibilities, among them the filtering of Internet searches.
