Slide 3 of 19
Notes:
So I go home and scratch my head
1. Concepts must be finely granular... Don’t want everything about “health”, or
even “health effects of food supplements” ... need “health effects of ginseng tea”
Acronyms -- ACP is American College of Physicians
also Associated Collegiate Press
also American Center for Physics
Names of Genes are a special problem in this regard - this is a big problem looming in biomedicine
Certain groups have special meanings to key phrases -
Zipf’s law says you shorten the words and phrases you use all the time
Thus “abortion” is used traditionally by physicians to mean a miscarriage,
but now it is used to mean a procedure...
3. If you are going to build something web-scale, must use a scalable methodology and produce a large scale vocabulary.