Text REtrieval Conference (TREC)
|
Organization Name: Technion- Israel Institute of Technology | Run ID: LARAg06pe5 |
Section 1.0 System Summary and Timing |
---|
Section 1.1 System Information |
Hardware Model Used for TREC Experiment:Power Mac G5 Quad System Use:SHARED Total Amount of Hard Disk Storage:233 Gb Total Amount of RAM:4096 MB Clock Rate of CPU:2500 MHz |
Section 1.2 System Comparisons |
Amount of developmental "Software Engineering":NONE List of features that are not present in the system, but would have been beneficial to have: List of features that are present in the system, and impacted its performance, but are not detailed within this form: |
Section 2.0 Construction of Indices, Knowledge Bases, and Other Data Structures |
---|
Length of the stopword list:430 words Type of Stemming:PORTER Controlled Vocabulary:NO Term weighting:YES
Phrase discovery:NO
Type of Spelling Correction:NONE Manually-Indexed Terms:NO Proper Noun Identification:NO Syntactic Parsing:NO Tokenizer:NO Word Sense Disambiguation:NO Other technique:YES Additional comments:Index of features generated for every document by a feature generator created from Wikipedia data. For more details, see the description in "Overcoming the Brittleness Bottleneck using Wikipedia: Enhancing Text Categorization with Encyclopedic Knowledge" by Gabrilovich and Markovitch in AAAI 2006. |
Section 3.0 Statistics on Data Structures Built from TREC Text |
---|
Section 3.1 First Data Structure |
Structure Type:INVERTED INDEX Type of other data structure used: Brief description of method using other data structure: Total storage used:6 Gb Total computer time to build:120 hours Automatic process:YES Manual hours required:hours Type of manual labor:NONE Term positions used:NO Only single terms used:NO Concepts (vs. single terms) represented:YES
Type of representation:mapping onto Wikipedia articles Auxilary files used:YES
Additional comments:The corpus documents were fed to the auxilary classifier, and the generated features were indexed in a standard inverted index. At retrieval time, features were generated for the query and this index was queried. |
Section 3.2 Second Data Structure |
Structure Type:INVERTED INDEX Type of other data structure used: Brief description of method using other data structure: Total storage used:22 Gb Total computer time to build:43 hours Automatic process:YES Manual hours required:hours Type of manual labor:NONE Term positions used:YES Only single terms used:YES Concepts (vs. single terms) represented:NO
Type of representation: Auxilary files used:NO
Additional comments:Standard bag of words index for the documents in corpus, used to create a conceptual model of the query. This index is queried in standard bag of words approach including query expansion, the top documents are processed to generate Wikipedia features, and the concepts shared by most documents are used as a query model. Then, a query is sent to the first inverted index, of Wikipedia concepts and final results are output. |
Section 3.3 Third Data Structure |
Structure Type: Type of other data structure used: Brief description of method using other data structure: Total storage used:Gb Total computer time to build:hours Automatic process: Manual hours required:hours Type of manual labor:NONE Term positions used: Only single terms used: Concepts (vs. single terms) represented:
Type of representation: Auxilary files used:
Additional comments: |
Section 4.0 Data Built from Sources Other than the Input Text |
---|
File type:OTHER Domain type:DOMAIN INDEPENDENT Total Storage:0.9 Gb Number of Concepts Represented:concepts Type of representation: Automatic or Manual:
Type of Manual Labor used: Additional comments: |
File is: Total Storage:Gb Number of Concepts Represented:concepts Type of representation: Additional comments: |
Section 5.0 Computer Searching |
---|
Average computer time to search (per query): CPU seconds |
Times broken down by component(s): |
Section 5.1 Searching Methods |
Vector space model: Probabilistic model: Cluster searching: N-gram matching: Boolean matching: Fuzzy logic: Free text scanning: Neural networks: Conceptual graphic matching: Other: Additional comments: |
Section 5.2 Factors in Ranking |
Term frequency: Inverse document frequency: Other term weights: Semantic closeness: Position in document: Syntactic clues: Proximity of terms: Information theoretic weights: Document length: Percentage of query terms which match: N-gram frequency: Word specificity: Word sense frequency: Cluster distance: Other: Additional comments: |
Section 6.0 Query Construction |
---|
Section 6.1 Automatically Built Queries for Ad-hoc Tasks |
Topic fields used: Average computer time to build query CPU seconds Term weighting (weights based on terms in topics): Phrase extraction from topics: Syntactic parsing of topics: Word sense disambiguation: Proper noun identification algorithm: Tokenizer: Expansion of queries using previously constructed data structures: Automatic addition of: |
Section 6.2 Manually Constructed Queries for Ad-hoc Tasks |
Topic fields used: Average time to build query? minutes Type of query builder: Tool used to build query: Method used in intial query construction? Total CPU time for all iterations: seconds Clock time from initial construction of query to completion of final query: minutes Average number of iterations: Average number of documents examined per iteration: Minimum number of iterations: Maximum number of iterations: The end of an iteration is determined by: Automatic term reweighting from relevant documents: Automatic query expansion from relevant documents: Other automatic methods: Manual methods used: |
Disclaimer: Contents of this online document are not necessarily the official views of, nor endorsed by the U.S. Government, the Department of Commerce, or NIST. |