國立高雄大學圖資館 |

Language: English

Back

透過交易新增來隱藏敏感的頻繁項目集 = Hiding Sensitive...

國立高雄大學資訊工程學系碩士班

透過交易新增來隱藏敏感的頻繁項目集 = Hiding Sensitive Frequent Itemsets through Transaction Addition

Record Type:	Language materials, printed : monographic
Paralel Title:	Hiding Sensitive Frequent Itemsets through Transaction Addition
Author:	張家境,
Secondary Intellectual Responsibility:	國立高雄大學
Place of Publication:	[高雄市]
Published:	撰者;
Year of Publication:	民100
Description:	55葉部份彩圖，表格 : 30公分;
Subject:	資料探勘
Subject:	Data mining
Online resource:	http://handle.ncl.edu.tw/11296/ndltd/92009588771079774289
Notes:	參考書目：葉43-45
Notes:	內容為英文
Summary:	資料探勘主要是用來從大量的資料中擷取出有用的知識，以用來幫助公司做出有效的決策。然而，在資料蒐集與數據傳播的過程中，卻可能引發隱私資料外洩風險。有關於個人、企業或組織的敏感資訊，在分享或是發佈時，應該以隱匿資訊的方式受到保護。也因此，隱私保護之探勘成為近年來重要的研究議題之一。在這篇論文裡，我們提出兩種透過新增交易資料的方法來隱藏敏感項目集。第一個方法以貪婪法為基礎，首先計算出可以用來完全的隱藏住敏感項目集的最大的新增交易筆數，再進行資料隱藏的動作。其中每筆新增交易資料的長度與新增的項目集取決於經驗法則裡的標準常態分佈，如此，也能進一步減少隱匿敏感項目時所產生的副作用。在第二個方法裡，我們提出一個演化式的隱私保護之探勘法，用來找出最適合的項目集以加入新增交易中，進一步得以隱藏住敏感項目集。此方法設計了三個變數來設計一個彈性的評估函數，並可根據使用者的喜好彈性地分配此三個變數的權重。此外，準大項目集的概念也被用來減少重新掃描資料庫的成本，以加快評估染色體的過程。最終，我們將透過實驗結果來評估演算法的效能。 Data mining technology is used to derive useful knowledge from large databases, thus being able to help make effective decision for companies. The process of data collection and dissemination may, however, cause the risk of privacy threats. Sensitive or personal information and knowledge of individuals, industries and organizations are required to be kept as private information before they are shared or published. Thus, the privacy-preserving data mining (PPDM) has become an important issue in recent years. In this thesis, two approaches for hiding sensitive itemsets by inserting new transactions are proposed. The first one is a greedy-based approach, which computes the maximal number of transactions to be inserted into the original database for totally hiding sensitive itemsets. The length of each inserted transaction and the itemsets within it are decided by some heuristic rules, such that the side effects from hiding sensitive itemsets could be reduced. The second one is an evolutionary privacy-preserving data mining method to find appropriate itemsets within inserted transactions for hiding sensitive itemsets. It uses a flexible evaluation function with three factors. Different weights are then assigned to three factors depending on users' preference. The concept of pre-large itemsets is also used in the GA-based approach to reduce the cost of rescanning databases, thus speeding up the evaluation process of chromosomes. Experimental results are finally performed to evaluate the performance of the proposed two approaches.

透過交易新增來隱藏敏感的頻繁項目集 = Hiding Sensitive Frequent Itemsets through Transaction Addition
張, 家境

透過交易新增來隱藏敏感的頻繁項目集 = Hiding Sensitive Frequent Itemsets through Transaction Addition / 張家境撰 - [高雄市] : 撰者, 民100. - 55葉 ; 部份彩圖，表格 ; 30公分.
參考書目：葉43-45內容為英文.
資料探勘Data mining

透過交易新增來隱藏敏感的頻繁項目集 = Hiding Sensitive Frequent Itemsets through Transaction Addition
LDR:04115nam0a2200289 450 001 300437
005 20170214095622.0
009 300437
010 0 $b 精裝
010 0 $b 平裝
100 $a 20170214d2011 k y0chiy50 e
101 1 $a eng $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 透過交易新增來隱藏敏感的頻繁項目集 $d Hiding Sensitive Frequent Itemsets through Transaction Addition $z eng $f 張家境撰
210 $a [高雄市] $c 撰者 $d 民100
215 0 $a 55葉 $c 部份彩圖，表格 $d 30公分
300 $a 參考書目：葉43-45
300 $a 內容為英文
314 $a 指導教授：洪宗貝博士、林浚瑋博士
328 $a 碩士論文--國立高雄大學資訊工程學系碩士班
330 $a 資料探勘主要是用來從大量的資料中擷取出有用的知識，以用來幫助公司做出有效的決策。然而，在資料蒐集與數據傳播的過程中，卻可能引發隱私資料外洩風險。有關於個人、企業或組織的敏感資訊，在分享或是發佈時，應該以隱匿資訊的方式受到保護。也因此，隱私保護之探勘成為近年來重要的研究議題之一。在這篇論文裡，我們提出兩種透過新增交易資料的方法來隱藏敏感項目集。第一個方法以貪婪法為基礎，首先計算出可以用來完全的隱藏住敏感項目集的最大的新增交易筆數，再進行資料隱藏的動作。其中每筆新增交易資料的長度與新增的項目集取決於經驗法則裡的標準常態分佈，如此，也能進一步減少隱匿敏感項目時所產生的副作用。在第二個方法裡，我們提出一個演化式的隱私保護之探勘法，用來找出最適合的項目集以加入新增交易中，進一步得以隱藏住敏感項目集。此方法設計了三個變數來設計一個彈性的評估函數，並可根據使用者的喜好彈性地分配此三個變數的權重。此外，準大項目集的概念也被用來減少重新掃描資料庫的成本，以加快評估染色體的過程。最終，我們將透過實驗結果來評估演算法的效能。 Data mining technology is used to derive useful knowledge from large databases, thus being able to help make effective decision for companies. The process of data collection and dissemination may, however, cause the risk of privacy threats. Sensitive or personal information and knowledge of individuals, industries and organizations are required to be kept as private information before they are shared or published. Thus, the privacy-preserving data mining (PPDM) has become an important issue in recent years. In this thesis, two approaches for hiding sensitive itemsets by inserting new transactions are proposed. The first one is a greedy-based approach, which computes the maximal number of transactions to be inserted into the original database for totally hiding sensitive itemsets. The length of each inserted transaction and the itemsets within it are decided by some heuristic rules, such that the side effects from hiding sensitive itemsets could be reduced. The second one is an evolutionary privacy-preserving data mining method to find appropriate itemsets within inserted transactions for hiding sensitive itemsets. It uses a flexible evaluation function with three factors. Different weights are then assigned to three factors depending on users' preference. The concept of pre-large itemsets is also used in the GA-based approach to reduce the cost of rescanning databases, thus speeding up the evaluation process of chromosomes. Experimental results are finally performed to evaluate the performance of the proposed two approaches.
510 1 $a Hiding Sensitive Frequent Itemsets through Transaction Addition $z eng
610 0 $a 資料探勘 $a 隱私保護 $a 貪婪法 $a 基因演算法 $a 準大項目集 $a 敏感項目
610 1 $a Data mining $a Privacy preserving $a Greedy approach $a Genetic algorithm $a Pre-large
681 $a 008M/0019 $b 464103 1134 $v 2007年版
700 1 $a 張 $b 家境 $4 撰 $3 518829
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 NUK $c 20111103 $g CCR
856 7 $z 電子資源 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/92009588771079774289