知識圖譜的自然語言查詢和關鍵詞查詢

章 緒論·································．1
1．1 研究背景及意義··················．1
1．2 研究現(xiàn)狀···························．3
1．2．1 知識圖譜自然語言查詢的
研究現(xiàn)狀························3
1．2．2 知識圖譜關鍵詞查詢的
研究現(xiàn)狀························4
1．3 存在的關鍵問題··················．5
1．4 研究內(nèi)容及創(chuàng)新點···············．7
1．5 本書組織結構·····················10
第2 章 自然語言查詢和關鍵詞查詢的
相關研究···························12
2．1 知識圖譜的自然語言查詢······12
2．1．1 語義關系識別················．12
2．1．2 自然語言聚集查詢···········．13
2．1．3 查詢映射·····················．14
2．1．4 多樣化的自然語言查詢······．15
2．2 知識圖譜的關鍵詞查詢·········16
2．2．1 模式圖························．16
2．2．2 多樣化的關鍵詞查詢········．17
2．3 兩種查詢共用的基礎技術······19
2．3．1 實體識別和實體鏈接········．19
2．3．2 解釋詞典·····················．19
2．4 眾包—輔助語義關系識別規(guī)則
挖掘·································20
2．5 知識圖譜的其他非結構化
查詢方式···························21
2．5．1 交互式查詢···················．21
2．5．2 實例查詢和樣例查詢········．22
第3 章 基于眾包的自然語言查詢中
語義關系識別規(guī)則挖掘·········23
3．1 問題描述及創(chuàng)新點···············23
3．2 眾包模型···························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 帶反饋的并行式迭代模型···．26
3．3 生成語義關系數(shù)據(jù)集和
依賴結構數(shù)據(jù)集··················27
3．3．1 眾包模型標記語義關系·····．27
3．3．2 Stanford Parser 生成依賴
結構··························．27
3．4 挖掘語義關聯(lián)規(guī)則···············28
3．4．1 挖掘語義關聯(lián)規(guī)則的算法···．28
3．4．2 算法MSAR 的復雜度·······．30
3．5 實驗結果及分析—眾包模型··31
3．5．1 實驗數(shù)據(jù)及評估標準········．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 帶反饋的并行式迭代模型···．35
3．6 實驗結果及分析—語義關聯(lián)
規(guī)則·································36
3．7 語義關系識別·····················38
3．7．1 語義關系識別的算法········．38
3．7．2 算法SRR 的復雜度··········．39
3．7．3 實驗結果及分析—語義關系
識別··························．39
3．8 本章小結···························40
第4 章 知識圖譜的自然語言聚集
查詢·································42
4．1 問題描述及創(chuàng)新點···············42
4．2 查詢流程···························45
4．3 查詢理解···························45
4．3．1 意圖解釋·····················．45
4．3．2 依賴結構分類················．46
4．3．3 從依賴結構中識別意圖解釋·．47
4．3．4 查詢理解的優(yōu)化·············．49
4．3．5 算法AIII 的復雜度··········．49
4．4 構建基本圖模式··················50
4．4．1 擴展的解釋詞典ED ·········．50
4．4．2 短語映射·····················．51
4．4．3 謂詞-類型鄰近集PT ·········．51
4．4．4 謂詞-謂詞鄰近集PP ·········．53
4．4．5 語義關系映射················．53
4．4．6 算法SRM 的復雜度·········．55
4．4．7 構建基本圖模式BGP········．56
4．4．8 算法BBGP 的復雜度········．57
4．5 將基本圖模式翻譯為
SPARQL 語句·····················58
4．5．1 數(shù)值型謂詞···················．58
4．5．2 翻譯基本圖模式·············．59
4．5．3 翻譯聚集·····················．59
4．5．4 算法TA 的復雜度···········．61
4．6 實驗結果及分析··················61
4．6．1 實驗數(shù)據(jù)集···················．61
4．6．2 各階段的優(yōu)化能力···········．61
4．6．3 算法的有效性················．63
4．6．4 與現(xiàn)有算法對比·············．65
4．6．5 回答錯誤的原因·············．66
4．7 相關問題及解決方案············67
4．8 本章小結···························69
第5 章 知識圖譜的自然語言查詢—
SPARQL 和關鍵詞··············70
5．1 問題描述及創(chuàng)新點···············70
5．2 查詢流程···························71
5．3 SPARQL 語句的生成過程······72
5．4 查詢分解···························73
5．4．1 查詢理解階段················．73
5．4．2 查詢映射階段················．74
5．4．3 執(zhí)行SPARQL 階段··········．74
5．4．4 查詢分解算法················．75
5．4．5 算法DQ 的復雜度···········．76
5．5 構建關鍵詞索引··················77
5．5．1 算法QUKI ···················．77
5．5．2 算法QUKI 的復雜度········．78
5．6 聚合SPARQL 結果子圖和
關鍵詞查詢························78
5．6．1 算法CSK ····················．78
5．6．2 算法CSK 的復雜度·········．80
5．7 實驗結果及分析··················81
5．7．1 算法的有效性················．81
5．7．2 回答正確的原因·············．83
5．7．3 回答錯誤的原因·············．84
5．7．4 以SPARQL 查詢?yōu)橹鲗У?br />優(yōu)勢··························．85
5．7．5 關鍵詞索引的效率···········．85
5．8 本章小結···························86
第6 章 知識圖譜的關鍵詞聚集查詢···88
6．1 問題描述及創(chuàng)新點···············88
6．2 查詢流程···························90
6．3 構建類型-謂詞圖·················90
6．3．1 關系提取·····················．90
6．3．2 關系標準化··················．91
6．3．3 類型-謂詞圖··················．92
6．4 查詢理解···························92
6．5 基于類型-謂詞圖構建
查詢圖······························94
6．5．1 查詢圖························．94
6．5．2 構建查詢圖··················．94
6．5．3 算法BQG 的復雜度·········．99
6．6 將查詢圖翻譯為SPARQL
語句·································99
6．6．1 數(shù)值型謂詞···················．99
6．6．2 翻譯一般路徑················．99
6．6．3 翻譯聚集·····················100
6．6．4 算法TQGS 的復雜度········102
6．7 實驗結果及分析···············．102
6．7．1 算法的有效性················102
6．7．2 輸入的可擴展性·············104
6．7．3 數(shù)據(jù)集的可擴展性···········106
6．7．4 組件的有效性················106
6．8 本章小結························．108
第7 章 總結與展望·····················．109
7．1 總結······························．109
7．2 展望······························．111
參考文獻····································．112