Bang
V. NGUYEN and Min-Yen KAN
13
Functional
Faceted Web Query Classification
2.
Actual Query Distribution
nManual classification
¨100 queries (limited, small indicative sample)
¨Randomly chosen
¨Reject non-English, offensive queries
¨From AllTheWebTM, 2002
nJudged by authors
¨Use only the query string and search results as evidence
¨Other data (e.g., clickthrough data) intentionally left out
nBroader impact
Experiment
Details
•Manually sample 100 English queries
•AllTheWebTM query
log
•12 polysemous queries removed
88:
Non ambiguous:
General and
Specific
Non ambiguous:
General and
Specific
12: Ambiguous