知识图谱的自然语言查询和关键词查询

章 绪论·································．1
1．1 研究背景及意义··················．1
1．2 研究现状···························．3
1．2．1 知识图谱自然语言查询的
研究现状························3
1．2．2 知识图谱关键词查询的
研究现状························4
1．3 存在的关键问题··················．5
1．4 研究内容及创新点···············．7
1．5 本书组织结构·····················10
第2 章 自然语言查询和关键词查询的
相关研究···························12
2．1 知识图谱的自然语言查询······12
2．1．1 语义关系识别················．12
2．1．2 自然语言聚集查询···········．13
2．1．3 查询映射·····················．14
2．1．4 多样化的自然语言查询······．15
2．2 知识图谱的关键词查询·········16
2．2．1 模式图························．16
2．2．2 多样化的关键词查询········．17
2．3 两种查询共用的基础技术······19
2．3．1 实体识别和实体链接········．19
2．3．2 解释词典·····················．19
2．4 众包—辅助语义关系识别规则
挖掘·································20
2．5 知识图谱的其他非结构化
查询方式···························21
2．5．1 交互式查询···················．21
2．5．2 实例查询和样例查询········．22
第3 章 基于众包的自然语言查询中
语义关系识别规则挖掘·········23
3．1 问题描述及创新点···············23
3．2 众包模型···························24
3．2．1 迭代模型和并行模型········．25
3．2．2 迭代式并行模型和
并行式迭代模型·············．25
3．2．3 带反馈的并行式迭代模型···．26
3．3 生成语义关系数据集和
依赖结构数据集··················27
3．3．1 众包模型标记语义关系·····．27
3．3．2 Stanford Parser 生成依赖
结构··························．27
3．4 挖掘语义关联规则···············28
3．4．1 挖掘语义关联规则的算法···．28
3．4．2 算法MSAR 的复杂度·······．30
3．5 实验结果及分析—众包模型··31
3．5．1 实验数据及评估标准········．31
3．5．2 迭代模型和并行模型········．32
3．5．3 迭代式并行模型和并行式
迭代模型·····················．33
3．5．4 带反馈的并行式迭代模型···．35
3．6 实验结果及分析—语义关联
规则·································36
3．7 语义关系识别·····················38
3．7．1 语义关系识别的算法········．38
3．7．2 算法SRR 的复杂度··········．39
3．7．3 实验结果及分析—语义关系
识别··························．39
3．8 本章小结···························40
第4 章 知识图谱的自然语言聚集
查询·································42
4．1 问题描述及创新点···············42
4．2 查询流程···························45
4．3 查询理解···························45
4．3．1 意图解释·····················．45
4．3．2 依赖结构分类················．46
4．3．3 从依赖结构中识别意图解释·．47
4．3．4 查询理解的优化·············．49
4．3．5 算法AIII 的复杂度··········．49
4．4 构建基本图模式··················50
4．4．1 扩展的解释词典ED ·········．50
4．4．2 短语映射·····················．51
4．4．3 谓词-类型邻近集PT ·········．51
4．4．4 谓词-谓词邻近集PP ·········．53
4．4．5 语义关系映射················．53
4．4．6 算法SRM 的复杂度·········．55
4．4．7 构建基本图模式BGP········．56
4．4．8 算法BBGP 的复杂度········．57
4．5 将基本图模式翻译为
SPARQL 语句·····················58
4．5．1 数值型谓词···················．58
4．5．2 翻译基本图模式·············．59
4．5．3 翻译聚集·····················．59
4．5．4 算法TA 的复杂度···········．61
4．6 实验结果及分析··················61
4．6．1 实验数据集···················．61
4．6．2 各阶段的优化能力···········．61
4．6．3 算法的有效性················．63
4．6．4 与现有算法对比·············．65
4．6．5 回答错误的原因·············．66
4．7 相关问题及解决方案············67
4．8 本章小结···························69
第5 章 知识图谱的自然语言查询—
SPARQL 和关键词··············70
5．1 问题描述及创新点···············70
5．2 查询流程···························71
5．3 SPARQL 语句的生成过程······72
5．4 查询分解···························73
5．4．1 查询理解阶段················．73
5．4．2 查询映射阶段················．74
5．4．3 执行SPARQL 阶段··········．74
5．4．4 查询分解算法················．75
5．4．5 算法DQ 的复杂度···········．76
5．5 构建关键词索引··················77
5．5．1 算法QUKI ···················．77
5．5．2 算法QUKI 的复杂度········．78
5．6 聚合SPARQL 结果子图和
关键词查询························78
5．6．1 算法CSK ····················．78
5．6．2 算法CSK 的复杂度·········．80
5．7 实验结果及分析··················81
5．7．1 算法的有效性················．81
5．7．2 回答正确的原因·············．83
5．7．3 回答错误的原因·············．84
5．7．4 以SPARQL 查询为主导的
优势··························．85
5．7．5 关键词索引的效率···········．85
5．8 本章小结···························86
第6 章 知识图谱的关键词聚集查询···88
6．1 问题描述及创新点···············88
6．2 查询流程···························90
6．3 构建类型-谓词图·················90
6．3．1 关系提取·····················．90
6．3．2 关系标准化··················．91
6．3．3 类型-谓词图··················．92
6．4 查询理解···························92
6．5 基于类型-谓词图构建
查询图······························94
6．5．1 查询图························．94
6．5．2 构建查询图··················．94
6．5．3 算法BQG 的复杂度·········．99
6．6 将查询图翻译为SPARQL
语句·································99
6．6．1 数值型谓词···················．99
6．6．2 翻译一般路径················．99
6．6．3 翻译聚集·····················100
6．6．4 算法TQGS 的复杂度········102
6．7 实验结果及分析···············．102
6．7．1 算法的有效性················102
6．7．2 输入的可扩展性·············104
6．7．3 数据集的可扩展性···········106
6．7．4 组件的有效性················106
6．8 本章小结························．108
第7 章 总结与展望·····················．109
7．1 总结······························．109
7．2 展望······························．111
参考文献····································．112