Min-Yen Kan and Danny C. C. Poo
Known Item Queries (JCDL 2005)
13/25
Query Classification Features
•Two examples:
•Hill Raymond Coding Theory – A First Course
•japan and cultural
•
•Distinguishing characteristics:
–Longer: cut-and-paste, copying from a reference
–Mixed Case
–Determiners: not present in unknown item or area searches
–Proper Nouns: specific subjects or author names
–Advanced Operators: title or author restrictions
–Keywords: indicative of a type of publication e.g., “journal”, “textbook”, “course”
–
•Use POS tagging to create a total of 16 features that embody these characteristics