發(fā)布時間：2018-03-09 04:13

  本文選題：信息檢索　切入點：語義分析　出處：《華中師范大學(xué)》2013年碩士論文　論文類型：學(xué)位論文

【摘要】：互聯(lián)網(wǎng)的蓬勃發(fā)展帶動了信息檢索技術(shù)的不斷成熟,搜索引擎已經(jīng)成為每個人都離不開的重要工具,人性化服務(wù)的時代背景也要求信息檢索技術(shù)向智能化發(fā)展。傳統(tǒng)的基于關(guān)鍵詞機械匹配的信息檢索方式已經(jīng)不能滿足科學(xué)研究和普通用戶的需求,因此基于語義的信息檢索成為當(dāng)前信息檢索研究的熱點,通過自然語言語句進行信息檢索已經(jīng)成為發(fā)展的趨勢。 面對自然語言查詢語句,目前的檢索系統(tǒng)往往不能夠精確的理解用戶的查詢請求；同時,在檢索的過程中,現(xiàn)有的技術(shù)往往將文檔中的語義信息丟棄。在對現(xiàn)有的信息檢索模型的分析研究下,我們發(fā)現(xiàn)單純的查詢語句處理和主題模型檢索并不能滿足用戶對檢索結(jié)果準(zhǔn)確率越來越高的要求。 分析現(xiàn)有的技術(shù)和研究成果,本文提出了一種基于雙層語義分析的文檔排序方法,分別通過查詢語句層次語義分析和文檔篇章層次語義分析,獲取信息檢索過程中所需的語義信息,從而提升搜索引擎性能。同時給出了基于雙層語義分析的全文檢索系統(tǒng)框架,該系統(tǒng)能夠在查詢語句層次上,對查詢語句進行語義處理和復(fù)述；在文檔篇章層次上,通過提取文檔中的潛在主題語義信息,用于優(yōu)化檢索結(jié)果。該方法通過結(jié)合查詢語句層次的語義信息和篇章層次語義信息,在向量空間模型的基礎(chǔ)上給出了基于雙層語義分析的文檔打分公式。 根據(jù)提出的基于雙層語義分析的全文檢索系統(tǒng)框架,設(shè)計并實現(xiàn)了原型系統(tǒng),并解決在系統(tǒng)實現(xiàn)的中的問題。通過對系統(tǒng)的實驗結(jié)果進行分析,驗證了這種基于雙層語義分析的全文檢索方法的有效性。
[Abstract]:With the rapid development of the Internet, the information retrieval technology is becoming more and more mature, and the search engine has become an important tool that everyone can not do without. The background of humanized service also requires the development of information retrieval technology to intelligence. The traditional information retrieval method based on keyword mechanical matching can no longer meet the needs of scientific research and ordinary users. Therefore, information retrieval based on semantics has become a hot topic in current information retrieval research, and information retrieval through natural language sentences has become a trend of development. In the face of natural language query statements, the current retrieval systems are often unable to accurately understand the user's query requests; at the same time, in the process of retrieval, The existing technologies often discard the semantic information in the document. We find that simple query processing and topic model retrieval can not meet the users' increasing demand for the accuracy of retrieval results. After analyzing the existing technology and research results, this paper proposes a method of document sorting based on two-layer semantic analysis, which is based on query sentence level semantic analysis and document text level semantic analysis, respectively. The semantic information needed in the process of information retrieval is obtained so as to improve the performance of search engine. At the same time, a framework of full-text retrieval system based on double-level semantic analysis is presented, which can be used in query sentence level. Semantic processing and retelling of query statements; at the document text level, by extracting semantic information about potential topics in the document, This method combines the semantic information of query sentence level and text level semantic information, and gives the document scoring formula based on two-layer semantic analysis on the basis of vector space model. According to the proposed framework of full-text retrieval system based on two-layer semantic analysis, the prototype system is designed and implemented, and the problems in the system implementation are solved. The effectiveness of this full-text retrieval method based on double-level semantic analysis is verified.
【學(xué)位授予單位】：華中師范大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2013
【分類號】：TP391.1

【參考文獻】

相關(guān)期刊論文 前1條

1 張琪玉;;網(wǎng)絡(luò)信息檢索工具增強關(guān)鍵詞檢索功能的措施[J];圖書館雜志;2001年01期

，

本文編號：1586930