Table 3.

Illustration of the Basilisk algorithm [41].

Input: Extraction Patterns in the Unannotated Corpus and their Extractions, Seed Lists Output: Updated List of Seeds

Procedure: $lexicon=\{seedwords\}$ for $i:=0$ 1. Score all extraction patterns with RlogF 2. $pattern pool$= top ranked $20+i$ patterns 3. $candidate word pool$ = extractions of patterns in $pattern pool$ 4. Score candidate words in $candidate word pool$ 5. Add top five candidate words to $lexicon$ 6. $i:=i+1$ 7. Go to Step 1.