Return to search

趨近一般化資料倉儲與資料探勘之效能評估模型 / Toward a More Generalized Benchmark Workload Model for Data Warehouse and Data Mining

隨著網際網路的發達以及資料庫技術的成熟,人們取得資料變得非常的容易,再加上許多網際網路的應用其實就是一個自動化的資料收集工具,資料量之大已幾近爆炸的程度。資料倉儲便是一種用來儲存大量歷史資料的資料庫,提供彙整或是統計的資訊,以提供決策使用的資訊技術。而資料探勘是從大量的資料當中把對於決策過程中有幫助的規則找出來,提供給管理人員做為決策的參考,開創新的商業契機。資料倉儲的效能表現對於使用者的工作效率有著深遠的影響。因此有些用以衡量與預測資料倉儲之效能與效率之工作量模式便孕育而生,一般稱之為績效評估工具,然而目前所公佈的一般資料倉儲績效評估工具是針對特定範圍領域建構出某些典型的領域規格,並沒有一個使用者需求導向的資料倉儲績效評估工具。在資料探勘方面,探勘結果的準確度比起資料探勘所花費的時間來得重要,目前卻沒有一個有效、使用者需求導向的工具來評估資料探勘結果的準確度。我們針對資料倉儲的效能評估以及資料探勘準確度評估,設計一個以使用者需求為導向的工作量模型,來評估資料倉儲與資料探勘工具。 / As growth of Internet and mature of database technology, people can get the data much easily than before. Many applications on Internet, in fact, are the tools of gather data automatically so that the amount of data is growing bigger and bigger. Data warehouse is one kind of database to store lots of historical data to offer statistical information for the information technology of decisions. Data mining is to find the useful rules for decisions from the amount of data to help the managers make decisions and create the new opportunities of business. The performance of data warehouse is import to user’s work efficiency. Therefore, there are some workload model arise to evaluate and predict the performance and efficiency of data warehouse called benchmark. However, the data warehouse specification announced these days are constructed to some typical domain specific, and the performance evaluation stand on synthetic workload. But, when the difference between the domain of data warehouse user applied and domain of performance evaluation tool is very large, the performance metric may different a lot to the result of benchmark tool. In data mining, the accuracy of mining result is important to business. The accuracy of mining result is more important than the time spend on data mining. However, there is no any useful tool to evaluate the accuracy of mining result and there is no any standard of performance criteria for data mining, either. We design a user requirement-oriented workload to evaluate performance of data warehouse and precision of data mining.

Identiferoai:union.ndltd.org:CHENGCHI/G0093356038
Creators邱士涵, Chiu,Shih-Han
Publisher國立政治大學
Source SetsNational Chengchi University Libraries
Language英文
Detected LanguageEnglish
Typetext
RightsCopyright © nccu library on behalf of the copyright holders

Page generated in 0.0062 seconds