Intelligent Google-based search prototype
Intelligent Google-based search prototype (IGBSP)
might do very intellectual search.
Functional problems to solve
Google itself returns very good
search results. But there are still some problems:
- Google doesn’t allow synonym
search. For instance, “clever person” query will not return pages with
“bright man” word combination.
- Google results are a bit raw.
There are only references to pages with answers. The answer is often
placed on a page together with irrelevant
PCnous is able to solve these
problems. That is:
- PCnous is able to make
preprocessing of the query and prepare several “synonym” queries based on
- PCnous is able to make post-processing:
separate relevant and irrelevant
Why is PCnous capable to solve the problems?
- PCnous can easily remember synonyms
(related concepts) for any specified concept (e.g. word
or phrase). That’s why PCnous
is able to prepare “synonym” queries based on initial query.
- PCnous can activate
concepts related to initial query (related concepts
will be put into the short memory) – let name
these related concepts “RCSM”. Next step – calculate correlation between “result
page of the Google’s search” and “RCSM”. The stronger is correlation
between “the result page” and “initial query” the higher is rating of “the
Additional feature of IGBSP
In addition, PCnous may summarize returned
results and rephrase the answer to initial user’s query.