1、中期考核,报告人:刘庆霞 南京大学计算机系Websoft研究组,报告提纲,课程情况科研工作概述关键词查询向结构化查询的转换实体摘要RDF数据集浏览展望,2018/8/5,2,,课程情况,已获学分: A类8个,B类9个,D类10个,2018/8/5,3,,科研工作概述,2018/8/5,4,,科研工作查询转换,背景普通用户习惯采用关键词查询传统关键词匹配不能利用数据结构化的优势问题关键词查询 “南京 川菜馆” 结构化查询 “城市:南京 AND 商户类型:川菜馆”,2018/8/5,5,,科研工作查询转换,主要方法关键词到属性的映射T候选查询的排序相关度打分:相似度重排序,2018/8/5,6,,
2、Top 3,科研工作查询转换,效果展示,2018/8/5,7,,科研工作实体摘要,背景实体描述的多与杂快速了解实体的关键性质的需求问题提取关键属性作为摘要摘要的精简性摘要的易读性,2018/8/5,8,,http:/dbpedia.org/resource/England p3/homepage uhttp:/www.direct.gov.uk/ p0/type p8/Country p0/type p8/PopulatedPlace p0/type p8/Place p0/type p2/Thing p0/type p5/Place p0/type p5/Country p1/label l
3、England p3/name lEngland p13/point l51.5 -0.11666666666666667 p12/long l-0.11666667 p12/lat l51.5 p0/type p6/_Feature p8/populationDensity 395.0 p8/areaTotal l1.3039554140297626E11 p8/areaTotal l1.30395E11 p8/leaderName p7/David_Cameron p8/leaderTitle Prime Minister of the United Kingdom p8/leaderNa
4、me p7/Elizabeth_II p8/leaderTitle lMonarch p8/demonym lEnglish p8/ethnicGroup p7/White_people p8/ethnicGroup p7/British_Asian p8/ethnicGroup p7/Black_British p8/ethnicGroup p7/Mixed_%28United_Kingdom_ethnicity_category%29 p8/ethnicGroup p7/British_Chinese p8/ethnicGroup p7/Classification_of_ethnicit
5、y_in_the_United_Kingdom p8/language p7/Cornish_language p8/regionalLanguage p7/Cornish_language p8/language p7/English_language p8/officialLanguage p7/English_language p8/largestCity p7/Capital_city p8/capital p7/London p8/motto l(French) p8/motto lGod and my right p8/currency p7/Pound_sterling p8/g
6、overnmentType p7/Devolution p8/ethnicGroupsInYear 2009 p8/anthem p7/National_anthem_of_England p8/anthem p7/God_Save_the_Queen subClass,Prefixes0http:/www.w3.org/1999/02/22-rdf-syntax-ns#1http:/www.w3.org/2000/01/rdf-schema#2http:/www.w3.org/2002/07/owl#3http:/ vs 长度带权0-1背包问题熟知度加权多背包问题避免语义重复,2018/8/
7、5,9,,初步效果http:/dbpedia.org/resource/England,科研工作实体摘要,2018/8/5,10,,SetCoverp0/type p8/Placep1/label lEnglandp3/name lEngland,MultiKnapsack + 1-Gram (W=200)p0/type p6/_Featurep13/point l51.5 -0.11666666666666667p0/type p8/Countryp8/governmentType p7/Devolutionp8/capital p7/Londonp3/name lEnglandp8/are
8、aTotal l1.3039554140297626E11p8/areaTotal l1.30395E11p8/language p7/Cornish_language,Knapsack (W=200)p12/lat l51.5p12/long l-0.11666667p0/type p8/Countryp0/type p5/Countryp1/label lEnglandp8/leaderTitle lMonarchp8/capital p7/Londonp3/name lEnglandp8/areaTotal l1.30395E11p8/demonym lEnglishp8/ethnicG
9、roup p7/Black_Britishp8/motto l(French),Knapsack + 1-Gram (W=200)p13/point l51.5 -0.11666666666666667p0/type p8/Countryp0/type p5/Countryp8/governmentType p7/Devolutionp8/capital p7/Londonp3/name lEnglandp8/areaTotal l1.3039554140297626E11p8/areaTotal l1.30395E11p8/language p7/Cornish_language,科研工作R
10、DF数据集浏览,背景RDF数据不易于人的阅读浏览查询构造不方便目标数据集显示数据集创建基于规则的查询的创建,2018/8/5,11,,科研工作RDF数据集浏览,RDF数据集显示概貌视图schema, statistics查询列表视图局部实例的图形化视图数据集创建简单合并方式由概念层出发的方式,2018/8/5,12,,科研工作RDF数据集浏览,基于规则的查询的构造多个规则构成一个查询自定义查询、原始关系均视为内涵型目标,可用于创建规则(以引用的方式)通过变量的类型检查来判断可用性,2018/8/5,13,,科研工作RDF数据集浏览,2018/8/5,14,,已有尝试 实例列表视图,科研工作RDF数据集浏览,2018/8/5,15,,已有尝试 关系列表视图,科研工作RDF数据集浏览,2018/8/5,16,,已有尝试 关系的编辑和创建,展望,2018/8/5,17,,谢谢!,2018/8/5,18,,