The Future of Medical Research

FSC729@yahoo.com · medicine forum beginner Joined: 24 May 2006 Posts: 2

Hello everyone I have an idea that could potentially lead to cures and
treatements for diseases that as of yet have no satisfactory treatement

or cure. My idea is the idea of generating new ideas, here is how it
works:

Massive banks of PC's connected together will process the following:

Make a giant computerized database of nearly all the most important
medical research papers, second each paper will be analyzed for word
patterns, word frequency, word usage, grammar, etc. A linguistic model
of each paper will be made and catagorized.

Then a new abstract will be generated by combining words, word pairs,
sentences, at random or following certain parameters and criteria as
set by the linguistic model. Now to check that the abstract might
contain new ideas, it will checked against the database of existing
papers, a linguistic metric will be used to gauge how "new" or
different the abstract is and to make sure the idea is new, an internet

search will be made to determine how many websites or papers contain
the same phrases, words, word usage, etc. To make sure the idea is new
random sentences will be chosen from the paper and checked using
google. First the entire sentence will be searched for then search for
the sentence leave one word from the search phrase, then leave off two
words from the search phrase, then three words and so on. Those
abstracts with the lowest average hits amongst all the searches and the

highest average linguistic metric distance will be collected and
viewed.

I take it for granted that many of papers may not make sense given that

they are put together at random, but the random jumble of words and
sentences might trigger a new idea in the hopefully competent person
reading the abstract.

If one is looking for a new idea closer to an existing idea, all one
needs to do is collect all the papers on that idea, and change the
wording and word usage in the paper very very slightly, so the idea
more or less makes sense yet might contain new possiblities. One can
make a list of words not often used and insert them into the paper.

Another idea is to make a database of the papers only in one particular

part of one field, then analyze the word usage and word pattern and
then make a statistical model to help predict the words and ideas that
might be used in papers in the near future. One can employ a genetic
algorithm to modify the statistical algorithm to make the predicted
papers correlate closely with papers in the near future. Basically you
are trying to predict the field in a systematic manner.

Now lets say you are looking for a new treatment or cure, so what you
can do is generate all the abstracts with the words that might appear
in the paper that will lead to this new treatment or cure, amongst all
those abstracts there may exist the abstract that will lead to the new
treatment or cure.

Of course there are many details to be worked out, but I believe that
if implemented correctly it will lead to new treatments and cures that
we would have never thought of.

John G.

betsyb · medicine forum beginner Joined: 16 Apr 2006 Posts: 9

<FSC729@yahoo.com> wrote in message
news:1148482815.342675.224660@g10g2000cwb.googlegroups.com...

FSC729@yahoo.com · medicine forum beginner Joined: 24 May 2006 Posts: 2

betsyb wrote:

Google