Query Processing Techniques for XML Databases

瀏覽統計 Email 通知 RSS Feed

簡歷

基本資料

Project title

Code/計畫編號

NSC95-2221-E019-094

Translated Name/計畫中文名

XML資訊庫查詢技術之研究

Project Coordinator/計畫主持人

Ya-Hui Chang

Funding Organization/主管機關

National Science and Technology Council

Co-Investigator(s)/共同執行人

余佳鍇
隴黔霏
黃治中
吳天佑
許奮輝

Department/Unit

Department of Computer Science and Engineering

Website

https://www.grb.gov.tw/search/planDetail?id=1311211

Year

2006

Start date/計畫起

01-08-2006

Expected Completion/計畫迄

01-07-2007

Bugetid/研究經費

431千元

ResearchField/研究領域

XML查詢技術
訂閱與刊載
索引技術
子序列比對
字串索引
查詢圖改寫
XML query processing
subscription/publishing
indexing technology
subsequence matching
string indices
query graph rewriting

本計畫提出一結合修改式的一般化霍福轉換(Generalized Hough Transformation, GHT)及視覺樣本之快速影像物件檢索方法，並據以研發大型影像資料庫的平行處理檢索方法及硬體製作，提供內嵌式系統即時影像檢索平台。快速地從大型多媒體資料庫系統檢索出符合使用者語意的物件不是一件容易的事，許多技術和理論都還有研究及改進的空間，例如一個影像物件在檢索時，該物件可能在影像中經過移動、旋轉或縮放大小，如此就增加了檢索的困難度。多媒體資料檢索包含資料偏移(Data Skew)及執行偏移(Execution Skew)雙重問題，平行處理架構必須適當，才能克服輸出入系統太慢或網路太窄，影響資料載入速度的問題；再者，使用一般化的循序執行電腦系統比對資料，無法善用媒體資料處理本身隱含平行處理的因子，為了解決這個問題，本計畫提出一平行物件比對的演算法，以利達到大型多媒體資料庫系統檢索即時處理的目標。本計畫為一三年期計畫，第一年，我們將設計一具有可高度平行化且可抵抗物件移動、旋轉或縮放大小的內容導向影像檢索方法；第二年，我們將著重在應用在大型影像資料庫檢索的平行處理演算法及分散式架構設計；第三年則期待能完成一軟硬體切割適當的內嵌式即時影像檢索系統及適合行動裝置的使用者回饋相關檢索方法設計。初步實驗結果驗證本計畫提出的方法的可行性。 XML is winning wide acceptance as the standard for a variety of information fission and fusion. Efficient query evaluation over XML documents is very important in this process since intermediate query results are also data trees for subsequent pipelining stages. In this project 「Query Processing Techniques for XML Databases」, we aim to investigate the subgraph-finding problems for both XML document trees and ad hoc XML data streams. Specifically, for a database of XML documents, we have developed a novel index technique to improve the current state-of-the-art alternatives. This approach utilizes multi-resolution region aggregates to allow more effective concurrent comparisons among groups of tag elements in a query. We would like to extend this work to handle more general query styles, such as OR twigs, dags (directed acyclic graphs), etc. To process continuous queries from subscribers, we plan to leverage our string index expertise for efficient aggregations of individual query graphs from user subscription profiles. Overall multi-pattern matching can be fast to compute. The preliminary result will be published in the ICDE 2006. This research shall strengthen the execution efficiency of XQuery clauses and the effective materialization of their support environments.

Keyword(s)

DSpace CRIS

Query Processing Techniques for XML Databases

基本資料

Description