Dissertation > Excellent graduate degree dissertation topics show
An Improved Relation-Based Information Retrieval Technology
Author: LiYan
Tutor: LiZhouJun
School: National University of Defense Science and Technology
Course: Computer Science and Technology
Keywords: Relationship - based information retrieval Relation extraction Query and Analysis Triplet structure
CLC: TP391.3
Type: Master's thesis
Year: 2007
Downloads: 22
Quote: 0
Read: Download Dissertation
Abstract
|
Research work show that the existing relationship - based information retrieval technology ( Relation - based Information Retrieval , hereinafter referred to as RIR) is superior to (term) or semantic (concept) - based IR technology , but there are still obvious defects that can not be clear relationship itself , can only express the concept of A , B , there is the concept of relationship . This paper presents an improved IR technology -IRIR (Improved Relation-based Information Retrieval), is to clarify the relationship between the value and attributes , the integration of the concept of relationship information based on the relationship between the expression of the triples (triple), The following matching method to obtain unknown information . Knowledge appear in the text is expressed as follows : R ( relation ) [ First Concept, Second Concept ] interrogative pronouns ( what) , the beginning of the query is expressed as : R ( relation ) [ First Concept, Unknown] for doubt adverb ( such as how) the beginning of the query expression : R (Unknown ) [ First Concept , Second Concept ] , when the known partial match consistent text query triples expression , the query unknown part of a value . Thus , not only can achieve similar the QA (query answer) functions can be completed accurate information retrieval . Drexel University DM Bioinformatics Lab developed based biomedical literature search engine ( 2004 version , referred to RIRS ) , we developed a technology and functional can achieve IRIR experiment the IR Engine - IRIRS of ( Improved Relation -Based IR System) , the system uses the UMLS and two authoritative WordNet ontology library to identify concepts and relationships , and the experimental results are satisfactory Dr. entrance examination in English reading comprehension test set and TREC 2004 Genomics Track on the test set , IRIRS retrieval accuracy of the text segment level MAPP (Mean average passage precision) from RIRS of 64.44 % to 74.28% : IRIRS than RIRS TREC overall document -level performance have improved significantly , the MAP from 21.71 % ( TREC ) , 37.58 % ( RIRS ) improve to 40.14% , 50 topic the average value of P @ 100 from 26.37% (TREC), 53.69% to 63.93% . This suggests that the improvement in IR applications based on relational information retrieval technology is very explore value .
|
Related Dissertations
- Research on Security-check Information System Based on Multi-information Fusion,V328.3
- Design and Realize of Public Transportation Inquiry System Based on WebGIS,TP311.52
- Design and Realization of GIS Component Based on COM Theory and Technology,P208
- Research on Several Key Technical Issues on Fine-Grained Sentiment Classification,TP391.1
- Research on Multi-view Commodity Ontology Learning,F224
- Efficient query and analysis of distributed spatial information system,P208
- The Research of Learning Algorithm in Reltion Extraction,TP18
- Research on Personalized WEB Search Based on Reference Document Model,TP391.3
- Research on Construct of Domain Ontology Based on Hierarchical Demand and Method of Semantic Annotation,TP391.1
- The Research for Named Entity Recognition and Relation Extraction in Text,TP391.1
- Research on Text Mining in Biomedical Literature,TP311.13
- Key Technologies for Distributed Virtual Harbour,U675.7
- Research on Semantic Relation Extraction between Named Entities,TP391.1
- Research on Associated Issues in Biomedical Text Mining Based on Discriminative Models,TP391.1
- Based on SVM and link analysis of protein relation extraction system,R319
- Research on Kernel Based Entity Relation Extraction,TP391.1
- Research on Relevance Feedback Based-Conceptual Graphs and Its Implementation,TP391.3
- Research of Named Entity Relation Extraction Method Based on Bootstrapping,TP391.1
- Implementation of Weather Query and Analysis System Based on 3DGIS,TP311.52
- The Research of Chinese Named Entity Recognition and Its Relation Extraction,TP391.4
CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Retrieval machine
© 2012 www.DissertationTopic.Net Mobile
|