Invited Talk 3

Invited Talk 3

We present several problems that appear in text mining, which stem from the way how many biomedical entities are named. Specifically, we show how this affects calculation of enrichment of individual terms and associated pairs of terms in the analyzed text, and how this affects the requirements for storage and run time during the text mining operations. We also present statistics of the promiscuous terms and effects of pruning such terms. The presentation concludes with several suggested guidelines.