|
REF
TAGGER (Artificial Intelligence Project)
|
|
 |
Ms. Alexander is researching neuro-degeneration in Alzheimer’s disease. Her work is
very critical since there is
no cure yet for Alzheimer’s.
|
|
|
She looks for contemporary work
and historical work in this area by sifting through hundreds of thousands of references
(on monograms,
research thesis,
conference papers,
journal articles
etc.,)
|
 |
|
| The effort required is high, but finding the appropriate
reference is like finding a needle in a haystack |
|
| The REFERENCE
TAGGER PROJECT aims to read unstructured references
(given below) as input into the system recognizes all the elements
like Author,
Article, Journal
Names, Year,
Publisher,
City etc., |
|
|
SAMPLE REFERENCES |
|
(a) Evins, D. A.; Nelsa, S. G.; J. Am. Chem. Soc. 1999, 159,
6452-6453.
|
| This is done using a combination of |
|
|
- Grammar developed using PERL
and regular expressions.
- Pattern recognition.
- Knowledge base built up into
indexed tables.
|
|
| Each element is recognized and tagged
as shown below. |
|
|
<parsed pref=0>0: <ref><jref><text:index>(a)</text:index>
<authors><author:last-first><lName>Evins</lName>,
<fName>D. A.</fName></author:last-first>; <author:last-first><lName>Nelsa</lName>,
<fName>S. G. </fName> </author:last-first></authors>;
<I><journal>J. Am. Chem. Soc</journal>.</I>
<B><year:single>1999</year:single></B>,
<I><volume>159</volume></I>, <fpage>6452</fpage>-<lpage>6453</lpage>.</jref></ref></parsed>
|
|
| The tagged references are fed into a
database indexing by Author,
Article,
Journal,
Editor,
Publisher, City
of Publishing,
Year….
So that any search by Ms. Alexander can be retrieved immediately
from the database. |
|
Next |
|
|
