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

基於樹狀結構的平均效用挖掘之維護方法 = Tree-based Main...

國立高雄大學電機工程學系碩士班

基於樹狀結構的平均效用挖掘之維護方法 = Tree-based Maintenance Approaches for Average-Utility Mining

紀錄類型:	書目-語言資料,印刷品 : 單行本
並列題名:	Tree-based Maintenance Approaches for Average-Utility Mining
作者:	陳柏諺,
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:	2012[民101]
面頁冊數:	66面圖，表 : 30公分;
標題:	資料探勘
標題:	data mining
電子資源:	http://handle.ncl.edu.tw/11296/ndltd/73045734333066540684
附註:	參考書目：面63-66
摘要註:	效益探勘為頻繁項目集探勘的延伸，它不僅考慮到項目集的頻率，並考慮項目集的成本、利潤或其他來自使用者喜好的考量。傳統上，一個項目集的效益是所有交易中項目集的效益的加總，並沒有考慮到項目集本身的長度。在過去有學者提出平均效益的衡量，並顯示其比原來的效益衡量可以得到更好的效益效果。此外也有學者設計出高平均效益樣式樹的結構用來幫助儲存一些相關的資料以便正確並有效的探勘出高平均效益項目集。由於在實際應用中，交易資料經常會有新增與刪除的情況，因此如何在動態的環境裡去維護高平均效益樣式樹以利效益挖掘變的相當重要。在本篇論文裡，我們分別針對交易資料的新增與刪除提出相關的高平均效益樣式樹維護演算法。我們提出的方法主要是基於快速更新演算法的維護概念，該方法最初是為頻繁項目集探勘而設計。根據項目在原始資料庫與新增資料或是刪除資料裡的平均效益高估值，我們可以將項目分成四個情況個別進行處理。從實驗結果，我們可以得知所提出的新增和刪除演算法，在執行的時間上都會比批次演算法來的更好，而在樹狀結構的節點個數比較上，也與批次方法的節點個數幾乎相同。 Utility mining is an extension of frequent-itemset mining. It does not only consider the frequency of itemsets, but also consider item cost, profit or other measures from user preference. Traditionally, the utility of an itemset is the summation of the utilities of the itemset in all the transactions regardless of its length. In the past, the average-utility measure was proposed and revealed a better utility effect of combining several items than the original utility measure. The high average-utility pattern tree (HAUP tree) structure was also designed to help keep some related information for efficiently and effectively mining high average-utility itemsets. Since transactions may be inserted or deleted in real applications, maintenance of the HAUP tree is thus very important in a dynamic environment. In this thesis, we thus propose two high average-utility pattern-tree maintenance algorithms for transaction insertion and deletion, respectively. The proposed algorithms are based on the concept of the Fast UPdated (FUP) algorithm, which was originally designed for mining frequent itemsets. Four cases are individually considered according to whether the average-utility upper-bound values of the items are high or low in the original database and in the inserted or deleted transactions. Experimental results show that the proposed maintenance algorithms run faster than the batch algorithm for handling inserted or deleted transactions and generate nearly the same tree structure as the batch algorithm does.

基於樹狀結構的平均效用挖掘之維護方法 = Tree-based Maintenance Approaches for Average-Utility Mining
陳, 柏諺

基於樹狀結構的平均效用挖掘之維護方法 = Tree-based Maintenance Approaches for Average-Utility Mining / 陳柏諺撰 - [高雄市] : 撰者, 2012[民101]. - 66面 ; 圖，表 ; 30公分.
參考書目：面63-66.
資料探勘data mining

基於樹狀結構的平均效用挖掘之維護方法 = Tree-based Maintenance Approaches for Average-Utility Mining
LDR:04047nam0a2200277 450 001 322945
005 20170214095237.0
009 322945
010 0 $b 精裝
010 0 $b 平裝
100 $a 20170214y2012 k y0chiy50 e
101 1 $a eng $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 基於樹狀結構的平均效用挖掘之維護方法 $d Tree-based Maintenance Approaches for Average-Utility Mining $z eng $f 陳柏諺撰
210 $a [高雄市] $c 撰者 $d 2012[民101]
215 0 $a 66面 $c 圖，表 $d 30公分
300 $a 參考書目：面63-66
314 $a 指導教授：洪宗貝博士、蕭培墉博士
328 $a 碩士論文--國立高雄大學電機工程學系碩士班
330 $a 效益探勘為頻繁項目集探勘的延伸，它不僅考慮到項目集的頻率，並考慮項目集的成本、利潤或其他來自使用者喜好的考量。傳統上，一個項目集的效益是所有交易中項目集的效益的加總，並沒有考慮到項目集本身的長度。在過去有學者提出平均效益的衡量，並顯示其比原來的效益衡量可以得到更好的效益效果。此外也有學者設計出高平均效益樣式樹的結構用來幫助儲存一些相關的資料以便正確並有效的探勘出高平均效益項目集。由於在實際應用中，交易資料經常會有新增與刪除的情況，因此如何在動態的環境裡去維護高平均效益樣式樹以利效益挖掘變的相當重要。在本篇論文裡，我們分別針對交易資料的新增與刪除提出相關的高平均效益樣式樹維護演算法。我們提出的方法主要是基於快速更新演算法的維護概念，該方法最初是為頻繁項目集探勘而設計。根據項目在原始資料庫與新增資料或是刪除資料裡的平均效益高估值，我們可以將項目分成四個情況個別進行處理。從實驗結果，我們可以得知所提出的新增和刪除演算法，在執行的時間上都會比批次演算法來的更好，而在樹狀結構的節點個數比較上，也與批次方法的節點個數幾乎相同。 Utility mining is an extension of frequent-itemset mining. It does not only consider the frequency of itemsets, but also consider item cost, profit or other measures from user preference. Traditionally, the utility of an itemset is the summation of the utilities of the itemset in all the transactions regardless of its length. In the past, the average-utility measure was proposed and revealed a better utility effect of combining several items than the original utility measure. The high average-utility pattern tree (HAUP tree) structure was also designed to help keep some related information for efficiently and effectively mining high average-utility itemsets. Since transactions may be inserted or deleted in real applications, maintenance of the HAUP tree is thus very important in a dynamic environment. In this thesis, we thus propose two high average-utility pattern-tree maintenance algorithms for transaction insertion and deletion, respectively. The proposed algorithms are based on the concept of the Fast UPdated (FUP) algorithm, which was originally designed for mining frequent itemsets. Four cases are individually considered according to whether the average-utility upper-bound values of the items are high or low in the original database and in the inserted or deleted transactions. Experimental results show that the proposed maintenance algorithms run faster than the batch algorithm for handling inserted or deleted transactions and generate nearly the same tree structure as the batch algorithm does.
510 1 $a Tree-based Maintenance Approaches for Average-Utility Mining $z eng
610 0 $a 資料探勘 $a 效益探勘 $a 平均效益 $a 快速更新演算法 $a 維護
610 1 $a data mining $a utility mining $a average utility $a FUP algorithm $a maintenance
681 $a 008M/0019 $b 542201 7540 $v 2007年版
700 1 $a 陳 $b 柏諺 $4 撰 $3 546968
712 0 2 $a 國立高雄大學 $b 電機工程學系碩士班 $3 166118
801 0 $a tw $b NUK $c 20120504 $g CCR
856 7 $z 電子資源 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/73045734333066540684