Please use this identifier to cite or link to this item: `http://hdl.handle.net/11455/98270`
DC FieldValueLanguage
dc.contributor呂瑞麟zh_TW
dc.contributor.author賴家慧zh_TW
dc.contributor.authorJia-Hui Laien_US
dc.contributor.other資訊管理學系所zh_TW
dc.date2018zh_TW
dc.date.accessioned2019-03-22T06:43:45Z-
dc.identifier.urihttp://hdl.handle.net/11455/98270-
dc.description.abstract本研究開發一個查詢DBpedia的自然語言問答系統，讓使用者的自然語言問題轉換為SPARQL結構查詢關聯資料集，而轉化過程先從問句中辨識出有用的實體，接著計算其數量並依據數量來選擇重要的實體，然後以重要實體為中心重新建樹並移除不需要的字詞而產生一個圖形(Graph)，接著根據子圖進行遍歷依序查詢DBpedia，然後會得到一組或多組的triple，再將triples結合其他額外的條件，最後產生完整的SPARQL語法並得到答案。 本研究我們採用QALD-7、QALD-6、QALD-5、QALD-4和QALD-3多語言問題測試資料來評估我們的方法，在完整測試集中，在QALD-7獲得平均精確度為0.19、平均召回率為0.23而平均F-measure為0.20；在QALD-6獲得平均精確度為0.31、平均召回率為0.54而平均F-measure為0.34；在QALD-5獲得平均精確度為0.33、平均召回率為0.43而平均F-measure為0.36；在QALD-4獲得平均精確度為0.25、平均召回率為0.34而平均F-measure為0.24；在QALD-3獲得平均精確度為0.40、平均召回率為0.50而平均F-measure為0.41，並在QALD-3中更進一步在問句類型中，我們採用平均F-measure值比較，在Aggregation得到0.32、List得到0.26而Other為0.48，實驗結果顯示我們方法能解決複雜問句。zh_TW
dc.description.abstractWe present a natural language question answering system that queries DBpedia to convert user's natural language question into a SPARQL structure queries over linked dataset. First, we identify useful entities from the question, then calculate the quantity and count the quantity to select an pivot from the question sentence, then re-create the tree with pivot and remove the stopwords to produce a graph. According to the sub-graph to traverse and query DBpedia sequentially, and then get one or more sets of triples. Then combine the triples with other additional conditions, finally generate the complete SPARQL syntax and get the answer. In this paper, we used QALD-7, QALD-6, QALD-5, QALD-4 and QALD-3 multilingual test dataset to evaluate our method. In the complete dataset, we achieve an average precision of 0.19, an average recall of 0.23 and an average F-measure of 0.20 on the QALD-7；an average precision of 0.31, an average recall of 0.54 and an average F-measure of 0.34 on the QALD-6；an average precision of 0.33, an average recall of 0.43 and an average F-measure of 0.36 on the QALD-5；an average precision of 0.25, an average recall of 0.34 and an average F-measure of 0.24 on the QALD-4；an average precision of 0.40, an average recall of 0.50 and an average F-measure of 0.41 on the QALD-3 and further in the query type question on QALD-3, we use an average F-measure to compare and achieve 0.32 on Aggregation, 0.26 on List and 0.48 on Other. Experimental results show that our method can solve complex questions.en_US
dc.description.tableofcontents目錄 摘要 i Abstract ii 目錄 iii 圖目錄 iv 表目錄 v 第1章 緒論 1 1.1 研究背景 1 1.2 研究動機 1 1.3 論文結構 2 第2章 文獻探討 3 2.1 基於鏈結資料的問答系統之挑戰(Challenges of Question Answering System over Linked Data) 3 2.2 基於鏈結資料的問答系統之方法(Approach of Question Answering System over Linked Data) 5 2.3 查詢生成(Query Generation) 6 2.4 評估指標 (Evaluation Metrics) 7 2.5 依賴關係樹(Dependency tree) 8 2.6 廣度優先搜尋法(Breadth First Search，BFS) 10 第3章 研究方法 11 3.1 系統架構概述 (Overall System Architecture) 11 3.2 自然語言前處理 (Natural Language Preprocessing) 12 3.3 實體辨識與分類 (Entity Detection and Classification) 13 3.4 問句分類與操作辨識(Query Classification and Operator Detection) 18 3.5 重要實體選擇(Pivot Selection) 21 3.6 問句處理(Question Processing) 23 第4章 實驗方法與結果分析 32 4.1 實驗環境與開發工具 32 4.2 資料集 32 4.3 實驗結果與分析 33 第5章 結論 38 參考文獻 39zh_TW
dc.language.isozh_TWzh_TW
dc.rights同意授權瀏覽/列印電子全文服務，2019-08-22起公開。zh_TW
dc.subject問答系統zh_TW
dc.subject自然語言查詢zh_TW
dc.subject重要實體選擇zh_TW
dc.subject圖形探索zh_TW
dc.subject廣度優先搜尋法zh_TW
dc.subjectNatural Language Queryen_US
dc.subjectPivot Selectionen_US
dc.subjectGraph Explorationen_US
dc.title在DBpedia中基於圖遍歷的問答系統zh_TW
dc.titleGraph Traversal-based Question Answering System over DBpediaen_US
dc.typethesis and dissertationen_US
dc.date.paperformatopenaccess2018-08-22zh_TW
dc.date.openaccess2019-08-22-
item.grantfulltextrestricted-
item.fulltextwith fulltext-
item.cerifentitytypePublications-
item.languageiso639-1zh_TW-
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.openairetypethesis and dissertation-
Appears in Collections:資訊管理學系