SP500215
NIST Special Publication 500-215: The Second Text REtrieval Conference (TREC-2)
Appendix B: System Features
Appendix
National Institute of Standards and Technology
D. K. Harman
HA QUERY CONSThUCTION - A[OCRerr]MA11CALLY BUILT QUERIES (ADHOC)
M NAME [OCRerr] GE [ CIARIT [ Lsl [OCRerr] HNC [OCRerr] NYU ] SYRACUSE [[OCRerr] QUEEt
Turned in various runs Natural languas
using everything but <title[OCRerr], <desc:
is
factors, definitions, Num, title, Topic, desc, nan, <nan>, <con
litle, deac, nan, con litle, desc, nan, con All nationality desc, nan con [OCRerr]clean:<desc
p.I.r Tim `0 B.Id 0.80 sec CPU time About 0.1 sec tsess than a Sec. 3 (average for c
<0.1 iec per query per query [4] < 1 sec. 3.0 (Stage 1) query)
1[OCRerr]d to C*mslim[OCRerr]
ighting
tppic terms) Yes, for ranlting only Yes [1] Yes Not used Yes Yes
traction Yes, for concepts only Not used Yes Yes No
Parsing _______________ Yes [2] __________ Not used Yes Partial No
Rule based syntactic
ae Disambiguation category
__________________ disambiguation Not used Yes No
Unknown words
[OCRerr]un Identification treated as proper
nouns Partial Yes No
Yes, part of the
proper noun
No Not used Yes identification al[OCRerr] No
Year
itterns? number Dates
Assoriations No Not used Yes No
tof Queries Yes (Stage 1); Yes
vious Constmcted (partially for one
clure Yes No Notused run-Stage2 Yes
Automatic term
ructure? expansion based on Term
corpus associations dusters [[OCRerr]I Word-pafr phra
C addition of Boolean
rs/Proximity Operators No Not used Yes (Stage 1) Yes
Similar terms in a
given topic
statement were
Boolean operators on identified manually
concepts portion of Syntactic for ineaact
[3] topics phrases matching of
concepts (Stage 2) None