Cost-Effective Conceptual Design

Cost-Effective Conceptual Design

ID:40843442

大小:534.40 KB

页数:25页

时间:2019-08-08

Cost-Effective Conceptual Design_第1页
Cost-Effective Conceptual Design_第2页
Cost-Effective Conceptual Design_第3页
Cost-Effective Conceptual Design_第4页
Cost-Effective Conceptual Design_第5页
资源描述:

《Cost-Effective Conceptual Design》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库

1、Cost-EffectiveConceptualDesignUsingTaxonomiesAliVakilianvakilian@mit.edu1,YodsawalaiChodpathumwanychodpa@illinois.edu2,ArashTermehchytermehca@oregonstate.edu3andAmirNayyerinayyeria@oregonstate.edu31CSAIL,DepartmentofEECS,MIT,Cambridge,MA,USA2DepartmentofComputerScience,UniversityofIllinois,Urb

2、ana,IL,USA3SchoolofEECS,OregonStateUniversity,Corvallis,OR,USAAbstractItisknownthatannotatingnamedentitiesinunstructuredandsemi-structureddatasetsbytheirconceptsimprovestheeffectivenessofansweringqueriesoverthesedatasets.Ideally,onewouldliketoannotateentitiesofallconceptsinagivendomaininadatas

3、et,however,ittakessubstantialtimeandcomputationalresourcestodosooveralargedataset.Aseveryenterprisehasalimitedbudgetoftimeorcomputationalresources,ithastoannotateasubsetofconceptsinagivendomainwhosecostsofannotationdonotexceedthebudget.Wecallsuchasubsetofconceptsaconceptualdesignfortheannotate

4、ddataset.Wefocusonfindingaconceptualdesignthatprovidesthemosteffectiveanswerstoqueriesovertheannotateddataset,i.e.,acost-effectiveconceptualdesign.Since,itisoftenlesstime-consumingandcostlytoannotatesmallnumberofgeneralconcepts,suchasperson,thanalargenumberofspecificconcepts,suchaspoliticiananda

5、rtist,weuseinformationonsuperclass/subclassrelationshipsbetweenconceptsintaxonomiestofindacost-effectiveconceptualdesign.Wequantifytheamountbywhichaconceptualdesignwithconceptsfromataxonomyimprovestheeffectivenessofansweringqueriesoveranannotateddataset.Ifthetaxonomyisatree,weprovethattheproble

6、misNP-hardandproposeanefficientapproximationalgorithmandanexactpseudo-polynomialtimealgorithmfortheproblem.Wefurtherprovethatifthetaxonomyisadirectedacyclicgraph,givensomegenerallyacceptedhypothesis,itisnotpossibletofindanyapproximationalgorithmwithreasonablysmallapproximationratioorapseudo-poly

7、nomialalgorithmfortheproblem.Ourempiricalstudyusingreal-worlddatasets,taxonomies,andqueryworkloadsshowsthatourframeworkeffectivelyquantifiestheamountbywhichaconceptualdesignimprovestheeffectivenessofansweringqueries.Italsoindicatesthatou

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。