互聯(lián)網(wǎng)、物聯(lián)網(wǎng)、云計(jì)算等信息技術(shù)的快速發(fā)展,政治、經(jīng)濟(jì)、軍事、工業(yè)等各個(gè)領(lǐng)域的傳統(tǒng)應(yīng)用開(kāi)始與之相結(jié)合,產(chǎn)生了比以往任何時(shí)候都要多的數(shù)據(jù)。同時(shí),智能移動(dòng)設(shè)備、傳感器、電子商務(wù)網(wǎng)站、社交網(wǎng)站等等數(shù)據(jù)來(lái)源每時(shí)每刻都在創(chuàng)造多種多樣的數(shù)據(jù)。面對(duì)如此大量的數(shù)據(jù),如何及時(shí)、有效地分析它們并從中提取出有價(jià)值的信息,是政府和企業(yè)亟待解決的問(wèn)題。例如,中國(guó)證券監(jiān)督管理委員會(huì)(CSRC)通過(guò)股票買家和賣家的交易價(jià)格和數(shù)量來(lái)判斷是否存在交易內(nèi)幕和炒家的操控;支付寶網(wǎng)絡(luò)科技公司通過(guò)分析支付寶用戶在網(wǎng)絡(luò)平臺(tái)上的消費(fèi)記錄獲取不同用戶的消費(fèi)習(xí)慣并制定相應(yīng)的市場(chǎng)策略;交通部以不同的時(shí)間間隔分析道路網(wǎng)絡(luò)的交通流量信息,并制定減少城市交通擁堵的政策。數(shù)據(jù)挖掘作為一種從大量數(shù)據(jù)中挖掘重要的、未知的、有潛在價(jià)值的模式的處理過(guò)程,被廣泛用于解決這類問(wèn)題。關(guān)聯(lián)規(guī)則挖掘(ARM)是數(shù)據(jù)挖掘的核心任務(wù)之一。然而,依賴支持度(value of support)提取模式的傳統(tǒng)方法并不能很好支持依據(jù)效用(value of utility)的模式提取。因此,效用模式挖掘,這是我們研究的主題,已經(jīng)出現(xiàn)了為了滿足這一需求。最近,在這一領(lǐng)域提出了許...

【文章頁(yè)數(shù)】：140 頁(yè)

【學(xué)位級(jí)別】：博士

【文章目錄】：
摘要
Abstract
Notation
Chapter 1 Introduction
    1.1 Background
    1.2 Motivations
    1.3 Contributions
    1.4 The Structure of Thesis
Chapter 2 Related Work and Preliminaries
    2.1 Related Work
        2.1.1 High Utility Itemset Mining Algorithms on static Database
        2.1.2 Top-k High Utility Itemset Mining on Static Database
        2.1.3 Closed High Utility litemset Mining on Static Database
        2.1.4 High Utility Itemset on Incremental Mining
        2.1.5 High Utility Itemset Mining on Data Streams
        2.1.6 Multiple Minimum Utility Threshold Methods
        2.1.7 Using Support and Utility Thresholds
        2.1.8 Variants Problems
    2.2 Preliminaries
        2.2.1 Definitions
Chapter 3 SSUP-Growth:A Novel Mining High Utility Itemset Algorithm with Single-Scan of Database
    3.1 Introduction
    3.2 Problem Statement
    3.3 Proposed Approach
        3.3.1 The Components of SSUP-Tree
        3.3.2 Building the SSUP-Tree
    3.4 Achieving UP-tree From SSUP-tree
        3.4.1 Update SSUP-tree with New Data
    3.5 Experimental Evaluation
        3.5.1 Settings
        3.5.2 Evaluation on Synthetic and Real Datasets
        3.5.3 Scalability
    3.6 Conclusion
Chapter 4 Mining High utility Itemset with An Improved MultipleMinimum Utility Based Approach
    4.1 Introduction
        4.1.1 Different From Previous Works
        4.1.2 Problem Statement
    4.2    The Proposed Approach
        4.2.1 Preliminaries
        4.2.2 The basic Idea
    4.3 HUI-MMU-UD Algorithm
    4.4 Experimental Result
        4.4.1 Pattern Analysis
        4.4.2 Runtime
        4.4.3 Memory Usage
    4.5 Conclusion
Chapter 5 LUIM: New Low Utility Itemset Mining Framework
    5.1 Introduction
        5.1.1 Motivation
    5.2 Problem statement
    5.3    Proposed Framework
        5.3.1 Low Utility Itemset Mining Framework
        5.3.2 Low Utility Generators Miner Algorithms
        5.3.3 Low Utility Itemset Mining Algorithm
    5.4 Performance evaluation
        5.4.1 Experimental Environment and Datasets
        5.4.2 Runtime
        5.4.3 Memory Usage
        5.4.4 Generated Items Comparison
        5.4.5 Discussion
    5.5 Conclusion
Chapter 6 FLUI-Growth:Frequent Low-Utility Itemsets Mining
    6.1 Introduction
    6.2 Problem Formulation
    6.3 The Proposed Method
        6.3.1 The components of FLUP-Tree
        6.3.2 Building LUP-Tree
        6.3.3 FLUI-Growth
    6.4 Experimental Result
    6.5 Conclusion
Chapter 7 Conclusion and future work
    7.1 Conclusion
    7.2 Future work
Bibliography
Acknowledgements
Publications



本文編號(hào)：3911104