Table 7.

Topic 200: keyword frequency rank

#	Term	Term count	Percentage	Rank
1	Lupus	869	23.80%	1
2	Diseas	753	20.70%	2
3	Activ	496	13.60%	3
4	Associ	476	13.10%	4
5	Serum	294	8.10%	5
6	High	274	7.50%	6
7	Protein	195	5.40%	7
8	Express	179	4.90%	8
9	Chang	108	3.00%	9

The original information of the keywords for Topic 200 is shown: (1) terms are extracted with stemming; (2) term counts are obtained from the first round retrieved passages, which are the top 1000 retrieved passages as the baseline; (3) the percentage is calculated based on 9 terms; (4) the rank depends on the term counts; (4) the parameters for this baseline are (k₁, b) = (2.0, 0.4) with the paragraph-based index.