國立高雄大學圖資館 |

語系: 繁體中文

說明(常見問題)

圖資館首頁

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 = A Gene...

國立高雄大學資訊工程學系碩士班

適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 = A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams

紀錄類型:	書目-語言資料,印刷品 : 單行本
並列題名:	A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams
作者:	魏佑恩,
其他團體作者:	國立高雄大學
出版地:	[高雄市]
出版者:	撰者;
出版年:	民99[2010]
面頁冊數:	76面圖，表 : 30公分;
標題:	間接關聯規則
標題:	Indirect association
電子資源:	http://handle.ncl.edu.tw/11296/ndltd/35599473994475676440
摘要註:	間接關聯規則其概念是藉由一組稱之為“中介子(Mediator)”的頻繁樣式的聯繫，將一組出現頻率較低的項目關聯起來。近年來，間接關聯規則被視為一種有用的規則型態，在許多應用上都可以揭示隱藏的有趣信息，例如推薦排名、找出替代項目或競爭項目、網頁瀏覽路徑、基因表現分析等。然而就我們所知，目前關於間接關聯探勘的研究的重點仍在於在靜態資料，幾乎沒有研究探討如何在資料串流中發現此類型的樣式。在本文中，將考慮於資料串流如何探勘間接關聯規則。不同於傳統在資料串流上探勘的研究，通常針對不同的資料串流模型分別討論如何進行探勘，我們提出了一個通用的探勘架構，此架構可以涵蓋目前最常使用的三種串流視窗模式，包括界標視窗模式、時間遞減視窗模式、滑動視窗模式等，且允許使用者自訂適合其需求的視窗模式。基於此通用架構，我們發展了兩個可有效探勘間接關聯規則的演算法。我們證明所提出的方法，GIAMS-IND與GIAMS-MED，皆保證不會產生錯誤的間接關聯規則，且其品質的誤差皆在一定的範圍內。我們並透過在合成以及實際的資料集上進行完整深入的實驗來驗證所提出的演算方法的效用與效能。 An indirect association refers to an infrequent itempair, each item of which is highly co-occurred with a frequent itemset called “mediator”. Although indirect associations have been recognized as powerful patterns in revealing interesting information hidden in many applications, such as recommendation ranking, substitute items or competitive items, and common web navigation path, gene expression, etc, all work conducted up to now has focused on mining indirect associations from static data; almost no work, to our knowledge, has investigated how to discover this type of patterns from streaming data. In this thesis, the problem of mining indirect associations from data streams is considered. Unlike contemporary research work on stream data mining that investigates the problem individually from different types of streaming models, we treat the problem in a generic way. We propose a generic framework that can encompass all classical streaming models, including landmark window model, time-fading window model, and sliding window model, and allows the flexibility of users specified window model through parameter settings to fit their needs. Based on this generic framework, we develop two efficient algorithms to fulfill the task of generating indirect associations in this context. We prove that the proposed two algorithms GIAMS-IND and GIAMS-MED can guarantee no false positive patterns and bounded error on the quality of the discovered indirect associations. Comprehensive experiments on both synthetic and real datasets under three widely used streaming models show the effectiveness and efficiency of the proposed algorithms.

適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 = A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams
魏, 佑恩

適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 = A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams / 魏佑恩撰 - [高雄市] : 撰者, 民99[2010]. - 76面 ; 圖，表 ; 30公分.
參考書目：面.
間接關聯規則Indirect association

適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 = A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams
LDR:04192nam0a2200277 450 001 272979
005 20170214100757.0
009 272979
010 0 $b 精裝
010 0 $b 平裝
100 $a 20170214y2010 k y0chiy05 b
101 1 $a eng $d chi
102 $a tw
105 $a ak am 000yy
200 1 $a 適用於串流資料中探勘間接關聯規則的通用型架構及演算方法 $d A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams $f 魏佑恩撰
210 $a [高雄市] $c 撰者 $d 民99[2010]
215 0 $a 76面 $c 圖，表 $d 30公分
314 $a 指導教授：林文揚博士
320 $a 參考書目：面
328 $a 碩士論文--國立高雄大學資訊工程學系碩士班
330 $a 間接關聯規則其概念是藉由一組稱之為“中介子(Mediator)”的頻繁樣式的聯繫，將一組出現頻率較低的項目關聯起來。近年來，間接關聯規則被視為一種有用的規則型態，在許多應用上都可以揭示隱藏的有趣信息，例如推薦排名、找出替代項目或競爭項目、網頁瀏覽路徑、基因表現分析等。然而就我們所知，目前關於間接關聯探勘的研究的重點仍在於在靜態資料，幾乎沒有研究探討如何在資料串流中發現此類型的樣式。在本文中，將考慮於資料串流如何探勘間接關聯規則。不同於傳統在資料串流上探勘的研究，通常針對不同的資料串流模型分別討論如何進行探勘，我們提出了一個通用的探勘架構，此架構可以涵蓋目前最常使用的三種串流視窗模式，包括界標視窗模式、時間遞減視窗模式、滑動視窗模式等，且允許使用者自訂適合其需求的視窗模式。基於此通用架構，我們發展了兩個可有效探勘間接關聯規則的演算法。我們證明所提出的方法，GIAMS-IND與GIAMS-MED，皆保證不會產生錯誤的間接關聯規則，且其品質的誤差皆在一定的範圍內。我們並透過在合成以及實際的資料集上進行完整深入的實驗來驗證所提出的演算方法的效用與效能。 An indirect association refers to an infrequent itempair, each item of which is highly co-occurred with a frequent itemset called “mediator”. Although indirect associations have been recognized as powerful patterns in revealing interesting information hidden in many applications, such as recommendation ranking, substitute items or competitive items, and common web navigation path, gene expression, etc, all work conducted up to now has focused on mining indirect associations from static data; almost no work, to our knowledge, has investigated how to discover this type of patterns from streaming data. In this thesis, the problem of mining indirect associations from data streams is considered. Unlike contemporary research work on stream data mining that investigates the problem individually from different types of streaming models, we treat the problem in a generic way. We propose a generic framework that can encompass all classical streaming models, including landmark window model, time-fading window model, and sliding window model, and allows the flexibility of users specified window model through parameter settings to fit their needs. Based on this generic framework, we develop two efficient algorithms to fulfill the task of generating indirect associations in this context. We prove that the proposed two algorithms GIAMS-IND and GIAMS-MED can guarantee no false positive patterns and bounded error on the quality of the discovered indirect associations. Comprehensive experiments on both synthetic and real datasets under three widely used streaming models show the effectiveness and efficiency of the proposed algorithms.
510 1 $a A Generic Framework and Algorithms for Mining Indirect Associations from Data Streams
610 0 $a 間接關聯規則 $a 資料串流探勘 $a 通用串流模型
610 1 $a Indirect association $a data stream mining $a generic streaming model
681 $a 008M/0019 $b 464103 2626 $v 2007年版
700 1 $a 魏 $b 佑恩 $4 撰 $3 483127
712 0 2 $a 國立高雄大學 $b 資訊工程學系碩士班 $3 353878
801 0 $a tw $b 國立高雄大學 $c 20101224 $g CCR
856 7 $z 電子資源 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/35599473994475676440