Difference between revisions of "Zhishi.links Results for OAEI 2011"
| (3 intermediate revisions by the same user not shown) | |||
| Line 3: | Line 3: | ||
|Subject=Ontology Alignment | |Subject=Ontology Alignment | ||
|Authors=Xing Niu, Shu Rong, Yunlong Zhang, Haofen Wang, | |Authors=Xing Niu, Shu Rong, Yunlong Zhang, Haofen Wang, | ||
| − | |Series= | + | |Series=OM |
|Year=2011 | |Year=2011 | ||
|Abstract=This report presents the results of Zhishi.links, a distributed instance matching system, for this year’s Ontology Alignment Evaluation Initiative (OAEI) campaign. We participate in Data Interlinking track (DI) of IM@OAEI2011. In this report, we briefly describe the architecture and matching strategies of Zhishi.links, followed by an analysis of the results. | |Abstract=This report presents the results of Zhishi.links, a distributed instance matching system, for this year’s Ontology Alignment Evaluation Initiative (OAEI) campaign. We participate in Data Interlinking track (DI) of IM@OAEI2011. In this report, we briefly describe the architecture and matching strategies of Zhishi.links, followed by an analysis of the results. | ||
|Conclusion=In this report, we have presented a brief description of Zhishi.links, an instance matching system. We have introduced the architecture of our system and specific techniques we used. Also, the results have been analyzed in detail and several guides for improvements have been proposed. | |Conclusion=In this report, we have presented a brief description of Zhishi.links, an instance matching system. We have introduced the architecture of our system and specific techniques we used. Also, the results have been analyzed in detail and several guides for improvements have been proposed. | ||
|Future work=We look forwards to build an instance matching system with better performance and higher stability in the future. | |Future work=We look forwards to build an instance matching system with better performance and higher stability in the future. | ||
| + | |Problem=Link Discovery | ||
|Approach=No data available now. | |Approach=No data available now. | ||
| − | + | |Implementation=Zhishi.links | |
| − | |Implementation= | + | |Evaluation=Accuracy Evaluation |
| − | |Evaluation= | ||
|PositiveAspects=No data available now. | |PositiveAspects=No data available now. | ||
|NegativeAspects=No data available now. | |NegativeAspects=No data available now. | ||
|Limitations=No data available now. | |Limitations=No data available now. | ||
| − | |Challenges= | + | |Challenges=– When it comes with the problem of homonyms, instance matching systems should exploit as much information as possible to enhance the discriminability of their matchers. Currently, subject to the fact that most descriptions given by New York Times are written in natural language, the performance of our semantic similarity calculator are constrained. We are considering more tests carrying out on datasets in different styles and designing a more robust system. |
| + | – In DI track, only three types of resources are involved. The special words in names, which are extracted as values of characteristic properties, are chosen manually. Some smarter strategies should be applied to accomplish this mission. | ||
|ProposesAlgorithm=No data available now. | |ProposesAlgorithm=No data available now. | ||
| + | |Model=Architectural | ||
|Methodology=No data available now. | |Methodology=No data available now. | ||
|Requirements=No data available now. | |Requirements=No data available now. | ||
| − | |Download-page= | + | |Download-page=http://apex.sjtu.edu.cn/apex_wiki/Zhishi.links |
|API=No data available now. | |API=No data available now. | ||
|InfoRepresentation=No data available now. | |InfoRepresentation=No data available now. | ||
| Line 34: | Line 36: | ||
|RelatedProblem=No data available now. | |RelatedProblem=No data available now. | ||
|Motivation=No data available now. | |Motivation=No data available now. | ||
| − | |ExperimentSetup= | + | |ExperimentSetup=Tests were carried out on a Hadoop computer cluster. Each node has a quad-core Intel Core 2 processor (4M Cache, 2.66 GHz), 8GB memory. The number of reduce tasks was set to 50. |
| − | |EvaluationMethod= | + | |EvaluationMethod=Utilize distributed MapReduce framework to adopt index-based pre-matching |
|Hypothesis=No data available now. | |Hypothesis=No data available now. | ||
|Description=No data available now. | |Description=No data available now. | ||
| − | |Dimensions= | + | |Dimensions=Accuracy |
| − | |Benchmark= | + | |Benchmark=DBpedia, Freebase, GeoNames |
|Results=No data available now. | |Results=No data available now. | ||
}} | }} | ||
Latest revision as of 13:00, 12 July 2018
| Zhishi.links Results for OAEI 2011 | |
|---|---|
Zhishi.links Results for OAEI 2011
| |
| Bibliographical Metadata | |
| Subject: | Ontology Alignment |
| Year: | 2011 |
| Authors: | Xing Niu, Shu Rong, Yunlong Zhang, Haofen Wang |
| Venue | OM |
| Content Metadata | |
| Problem: | Link Discovery |
| Approach: | No data available now. |
| Implementation: | Zhishi.links |
| Evaluation: | Accuracy Evaluation |
Abstract
This report presents the results of Zhishi.links, a distributed instance matching system, for this year’s Ontology Alignment Evaluation Initiative (OAEI) campaign. We participate in Data Interlinking track (DI) of IM@OAEI2011. In this report, we briefly describe the architecture and matching strategies of Zhishi.links, followed by an analysis of the results.
Conclusion
In this report, we have presented a brief description of Zhishi.links, an instance matching system. We have introduced the architecture of our system and specific techniques we used. Also, the results have been analyzed in detail and several guides for improvements have been proposed.
Future work
We look forwards to build an instance matching system with better performance and higher stability in the future.
Approach
Positive Aspects: No data available now.
Negative Aspects: No data available now.
Limitations: No data available now.
Challenges: – When it comes with the problem of homonyms, instance matching systems should exploit as much information as possible to enhance the discriminability of their matchers. Currently, subject to the fact that most descriptions given by New York Times are written in natural language, the performance of our semantic similarity calculator are constrained. We are considering more tests carrying out on datasets in different styles and designing a more robust system. – In DI track, only three types of resources are involved. The special words in names, which are extracted as values of characteristic properties, are chosen manually. Some smarter strategies should be applied to accomplish this mission.
Proposes Algorithm: No data available now.
Methodology: No data available now.
Requirements: No data available now.
Limitations: No data available now.
Implementations
Download-page: http://apex.sjtu.edu.cn/apex wiki/Zhishi.links
Access API: No data available now.
Information Representation: No data available now.
Data Catalogue: {{{Catalogue}}}
Runs on OS: No data available now.
Vendor: No data available now.
Uses Framework: No data available now.
Has Documentation URL: No data available now.
Programming Language: No data available now.
Version: No data available now.
Platform: No data available now.
Toolbox: No data available now.
GUI: No
Research Problem
Subproblem of: No data available now.
RelatedProblem: No data available now.
Motivation: No data available now.
Evaluation
Experiment Setup: Tests were carried out on a Hadoop computer cluster. Each node has a quad-core Intel Core 2 processor (4M Cache, 2.66 GHz), 8GB memory. The number of reduce tasks was set to 50.
Evaluation Method : Utilize distributed MapReduce framework to adopt index-based pre-matching
Hypothesis: No data available now.
Description: No data available now.
Dimensions: Accuracy
Benchmark used: DBpedia, Freebase, GeoNames
Results: No data available now.