知识图谱构建的质量控制

所需积分/C币:48 2018-09-08 00:23:06 1.87MB PDF
收藏 收藏
举报

复旦大学知识图谱培训ppt7:知识图谱构建的质量控制(Data Quality Issues in Constructing Knowledge Graph)
3 988EB Data Explosion! (2010 BleB (2006 by IDC) 情 32 billion GB 2003 导 多 20 billion GB 2002 40,000 BCE 12.5 billion GB 2001 1970 6.2 billion GB 1980 2000 Source: Horison Information Strategies, cited from Storage New Game New Rules, p. 34 1990 www.horison.com 2000 Knowledge Explosion? Indexing, search, data mnIng, query data volume optimization, data streams, web scale data management, information extraction and understanding, data-driven research data diversity parallel processing Data integration federated databases domain-specific query knowledge processing, ontology, semantic web, dataspace data quality DQ Problems in DBLP Wei Wang 16 Tao wang 18 Jun Zhang 21 5 Wei li: 27 Lei Wang 30 a Polyseme: 10+ different "Wei Wang Michael Wagner: 5 口 Synonyms:" Pei lee"and“ Pei li” Jim smith 3 dblp PeiLi & s甲 computer science bibliography Home Persons Pei Lee tS甲 [+ Other persons with a similar name 4 Search dblp Home persons [-1 Journal Articles o Ho [+ other persons with a similar name O Author Nsults是 2015 [- Conference and Workshop Papers o 19]自 Teng Li, Jian Mac, Exact matches Rating cloud stor 2014 Wei Wang Wei war ■18]目 Jinxin Zhang, Chad■[5]目见唆 3-D simulation st Pei Lee, Laks V.S. Lakshmanan, Evangelos E Milios National Univers fy of Singapore Wei Wang ooo Reliability 55(8): 1 CAST: A Context-Aware Story-Teller for Streaming Social Co College of Nano ale Science, University at Albany/Purdue Univer■17目丛 Haibin Duan,Pei■c4]目& Pei Lee, Laks V.s. Lakshmanan, Evangelos E. Milios Wei Wang ooo Interactive learn Incremental cluster evolution tracking from highly dynam School of Life Sd ence, Fudan University, China 015 Wei Wang [c3]&< Pei Lee, Laks V S Lakshmanan, Mitul Tiwari, Sam Shah: Center for Engi leering and Scientific Computation, Zhejiang Unive 2014 show all Modeling impression discounting in large-scale recommen ■016]目息 Yingwen Chen,Mi ikely mato Empirical studyc“013 2014:180(2014) [c2]E&s Pei Lee, Laks V S Lakshmanan, Evangelos EMilios ng010 UCLA University of North Carolina at Chapel Hill ■[15]自 Pei li, Yunchuan s Keysee: supporting keyword search on evolving events in s u Wei Wang oos Modeling and pel 2012 Fudan University, shanghai, China Commuinication Sy Weidong Wang c1自零 Pei lee, Laks v.S. Lakshmanar,efey‰uYu: Wei-Fan Wang aka: Weifan Wang On Top-k Structural similarity Search. ICDE 2012: 774-785 show a 357 matches Difficult Names in Google search 6 488941 britney spears 29 orient spears g brittany spears 5 barney spear= 3yit1ya≌eax 40134 brittany spears 29 brittany spears s britanay spears 5 britney spears 3 britmeny spe spears 36315 brittney spears 29 brittany spears 9 britinany spears 5 brotny spears 3 britney spears 2 brittany spears 24342 britan 2s上 tiney spear 5 bruteny apears 2上x1工 toney spears 7331 britny spears 26 birttney spears s britney spears 5 btiyney apears 3 britney spears 2 britain spears 6633 briteny spears 26bxe工 chey spears 9 britney spears MEleNey s÷arB britney spears 2上 ClAne岛wea 25s如 ritter 26 trinity spears s britney spear s riney apears 3 britnetty apears 2 britaneny spears 1a07 briney sy母a1多 26 britenay spears 9 briny spears 5 britney spears 3如ritn色 r spears 21ri七 ana eAi 1635y1 26 britney speare g brittney spears 4 bitty spear. 3如x1 -neyxxx spears 2上y1 cann spears 1479 brintey spears 26 brittan spears 9 briny apears 4 britney spears 3 britnity spears 2 britanna spears 1479 britanny spears 26 brittle spears Yony spears brandy spears 3 britney spears 2 britannie spears 1338 britiny spears 26 btittany spears s whitney spears 4 britney 3 ritner spea工g 2 britannt spea工9 1211 bitnet. spears 24 britney spears 8 birtiny spears 4如 reatiny8 ears 3 britterny spears 2 hritannu spears 1096上x1 LIney spears 24上 artery sPears bithney spears 4 breetney spears 3如 I1=cheey spear 2 bricanyl spears 991 brit 24上 rIgney spea 8 brittany spear 3 brittney spea 2 brit 991 britney spears 2 brintiny spears 8 breitny spears 4 britney spears 3 brittney spears 2 briteeny spears 日11r1 caney spears 2上 PicarDy母ax多 上 etery ears brittany spears 3 britten spears 2上 ritenany spear= 811 brtiney spears 24 britennv spears 8 righty spears 4 brieteny spears 3 britney spears 2 britenet spears 664 birney 24 britini spears 8 brintay spears 4 briery spears 3 britney spears 2 briteniy spears 664 britney spears 24 britney spears 8 brinkley ears 4 briitny spears 3加 roEenv spears 2 britenys spears 664 britney spears 24 brittani spears 8 britney spears brittany spears 3 brtaney spears 2brit⊥aney8 sears 601 bitney spears 24上11tt1号eay 日上 ricans spears 4 brine spear 3 brtiiany spears 2 ritin spears 60- brinty spears 21上 britney gears 8 briley spears 4 orintenev spea工 3 brtinay apears 2 britinary apears 544 brittaney spears 1 birtany spears 8 britney spears 4 brintne spears 3 brtinney spears 2 briery spears 544上x1 Chay spears 21上 cery spearE 日上 PItney spear= a britaby spears 3 brtitany spears 2上r1tran= y spears 3∈4 ratey8eax 21 bratley spears 8x⊥ Enty spears 3 brtiteny spears 2 britnat spears 364 brittiny spears 21 britani spears 8 bittner spears britainey spears 3 brunet spears 2 britney spears 329 britney spears 211七am1eeaY 8 brittany spears 4如it1 nIe sceaL 3 brytiny spears 2 britney spears 269 britney spears 21 briteany spears britney spears 4 britinney spears 3 boney spears 2 britney spears 269 britneys spears 21如x亡亡γ三口母豆 7 birney spears 4 romney spears 3 brittney spears 2上 Michener spear 224 britne spears 21上xi七 Clay sea工g 7 biteney spears 4 britne 工 =chev steaL 2上 riney69peax9 244 britney spears 21 pItal望 DEars 7上 icIly spears britnel spears 3 britney spears Yesa工s 220上 reamey === 21上xt1aYsF母as 7 breateny spear= a britney spears 2a工 cany spears 2 britney spears 220 britiany spears 19 birney spear 7 riant spea工g 生 r1tnewy spears 2x⊥ tney spears 2 britney spears 199 britney spears 19 britney spears 7 brintye spears 4 britney spears 2 baine : y spears 2 britneyyy spears 163 britney spears 19上r1 chae spears 7x1÷1 anny spear 4 brittany spears 2≥xi=ny2 ears 2上 richey spears 147 breathy spears i9 britnee a工三 7如rit1 y spears 4 brittery spears 2 brittany spears 2 britney spears 147bx1 LtIn=r s母a8 19上 reAny sp=aI 7如 1he spears 4 brittney spears 2上 el-any spears 2 brienne spears 孔 7 britney spears 4 brittney apears 2 beitny spears 2 britnu spears 147 britney spears 1g brittney spears 7如 ridley8 Cears brittnau spears 2如 raney8 pears 2 britney spears 47上 ruthey spe 17 birtny spears 7上x1-nay 4上 1=cheny spears 2 berth spears 2 brittany spears 33 brittney 17 brieny spears 7 brittian spears 4 brittnye spears 2如 edney spear 22xt工enY9Peax8 33 britney spears 17 brintty spears briny spears 4 brittany spears 2 beany spears 2 briery spears 21 dittany spears 17上 r1-ay3earg 7 brittany spears 4 britney spears 2 bariey spear 2 brittany spears Another Example with KBs 7 CIDe Name+ Address中 Citye Sexe 114 张三心 邯郸路20号计算机楼52室上海口 244 李四4 鄞秦路918号?号楼10室“宁波 CNOP Namee Gender Addresse PhoneFaxe 24王五 F 杭州市朝晖二区555号2308窒3100124 0571-88480666/4 0571-87074789 483李四M宁波市郵路978号7号楼2331207487189 NO+ Name+ Gender Address 中 Pone+ Fax+ CID+ Cno+ 14张三F邯郸路220号上 计算机楼527海 室 24李四M鄄奏路978号宁3150124057487074789 2444934 702室波 3王五F朝二区55号杭|31001241571:84806605714424 2308室州 884806674 口 Different schemas:e.g,“Sex”-“ Gender”," Phone/Fax”“ Phone"+"Fax” Inconsistency values:eg,“0/1”“午F/M” a Missing valves Six DQ Dimensions 8 Figure 1: Data Quality Dimensions Is all necessar Are all data valuas within the ata present lue domains specified by the business? Does dala rolled the real-world objects or a ks the data Data Quality Accuracy verifiable gourDe? available at the time needed Timeliness Is data consistent Are the relations between betwen systems? Do entities and attributes duplicate records east? consistent? within tables and between? The Taxonomy of dQ problems 9 Data quality Problems 数据质量问题 Single Data Source Multiple Data Source 单数据源问题 多数据源问题 Schema Instance Schema Instance 模式相关 实例相关 莫式相关 实例相关 (缺乏完整性约束, (数据输入错误)(不同的数据模型 (矛盾的或不一致 粗劣的设计模式) 一拼写错误 和模式设计) 的数据) Typos 唯一值 冗余/重复 命名冲突 不一致的聚集层级 Uniqueness Redundancy Naming Conflicts Inconsistency Aggregation Hierarchies 参考完整性 矛盾的数据 结构冲突 不一致的时间点 Referential integrit Conflicts Structural conflicts Inconsistency Time Points Computational Data Quality Problems 10 ○ Data Integration OData Imputation Schema Mapping O Data provenance ○ Record matching ● Data Uncertaint ● Data Cleaning ● Data constraints

...展开详情
试读 59P 知识图谱构建的质量控制
立即下载 低至0.43元/次 身份认证VIP会员低至7折
一个资源只可评论一次,评论内容不能少于5个字
alto1394 Knowledge Graph 知识图谱构建的质量控制 苏州大学 先进数据分析研究中心 李直旭
2018-11-06
回复
关注 私信 TA的资源
上传资源赚积分,得勋章
最新推荐
知识图谱构建的质量控制 48积分/C币 立即下载
1/59
知识图谱构建的质量控制第1页
知识图谱构建的质量控制第2页
知识图谱构建的质量控制第3页
知识图谱构建的质量控制第4页
知识图谱构建的质量控制第5页
知识图谱构建的质量控制第6页
知识图谱构建的质量控制第7页
知识图谱构建的质量控制第8页
知识图谱构建的质量控制第9页
知识图谱构建的质量控制第10页
知识图谱构建的质量控制第11页
知识图谱构建的质量控制第12页

试读结束, 可继续读6页

48积分/C币 立即下载 >