SP500207 CLARIT TREC Design, Experiments, and Results chapter D. Evans R. Lefferts G. Grefenstette S. Handerson W. Hersh A. Archbold National Institute of Standards and Technology Donna K. Harman Formating Text Prep T NLP Morph I Parsing I Candidate NPs Term/Doc Statistics [OCRerr]NP Scoring - [OCRerr] Matching Thesaurus Scoring - Filtering Lexicon `Core' Lex (100,000 items) Optional Su[OCRerr]Domain Lex `Heuristic' Grammar `Simplex' NP Optional "Complex" NP Optional "Full Sentence" Constituents General Set of Terms "Exact" "Novel" "General" Specific Set of Terms "1s[OCRerr]Order" Thesaurus: Flat List of Terms, Implicit Compositional, Hierarchical Structure Figure 1: `St[OCRerr]d&d' CLARIT Indexing Overview Formating [~w Text Prep NLP Scoring I Morph I Parsing I Candidate NPs Lexicon `Core' Lex (100,000 items) `Heuristic' Grammar `Simplex' NP Term/Doc_________ I { NPs & Words{ General Set of Terms Statistics Figure 2: Modified CLARIT Indexing in TREC 253