English  |  正體中文  |  简体中文  |  Items with full text/Total items : 51491/86611 (59%)
Visitors : 8248273      Online Users : 142
RC Version 7.0 © Powered By DSPACE, MIT. Enhanced by NTU Library & TKU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    Please use this identifier to cite or link to this item: http://tkuir.lib.tku.edu.tw:8080/dspace/handle/987654321/94450


    Title: MOGA-based multi-level genetic-fuzzy mining techniques
    Other Titles: 多目標為基礎的階層式遺傳模糊探勘技術
    Authors: 何吉軒;He, Ji-Syuan
    Contributors: 淡江大學資訊工程學系碩士班
    陳俊豪;Chen, Chun-Hao
    Keywords: 資料探勘;模糊集合;多目標遺傳演算法;分類階層;分群技術;隸屬函數;利潤模糊商品集;模糊關聯規則;data mining;Fuzzy set;multi-objective genetic algorithm;taxonomy;clustering technique;membership function;utility fuzzy itemset;fuzzy association rule
    Date: 2013
    Issue Date: 2014-01-23 14:39:35 (UTC+8)
    Abstract: 現實世界的交易資料通常包含購買數量,故許多因應此類型交易資料的模糊探勘方法被提出並用來挖掘模糊關聯規則。因為隸屬函數對最終探勘結果有重大影響,所以許多遺傳模糊探勘方法進一步被提出用來同時探勘隸屬函數與模糊關聯規則。然而,大部分的方法都專注於單一階層探勘並且只考慮一個目標函數。有鑑於此,本論文提出兩個方法來探勘柏拉圖集合(隸屬函數)並挖掘多階層模糊關聯規則,分別為多目標多階層遺傳模糊探勘方法(MOMLGFM)與兩階段多目標遺傳模糊探勘方法(TMOGFM)。
    在第一個方法(MOMLGFM),首先會根據給定的分類階層將商品類別的隸屬函數編成染色體。方法中考慮兩個目標函數。第一個目標函數是不同階層探勘出的資訊總和,第二個目標函數是染色體中隸屬函數的適合度。接著,每個個體的適性函數值則由這兩個目標函數計算而得。在演化流程完成後,多種隸屬函數則可根據決策者的喜好用來探勘多階層模糊關聯規則。
    然而,MOMLGFM找出柏拉圖集合後,決策者會有難以從中挑選出適合的隸屬函數進行規則探勘的困擾。所以在第二個方法(TMOGFM),我們以第一個演算法為基礎提出兩階段多目標模糊探勘演算法來幫助決策者選擇恰當的隸屬函數。在第一階段,使用MOMLGFM挖掘隸屬函數的柏拉圖集合。在第二階段,依據設計的規則或利潤導向分群屬性,透過分群技術將柏拉圖解分成不同群組並找出群組代表解。之後,依據決策者的喜好,每個群組內所選出的代表解則能用來探勘模糊關聯規則或利潤模糊商品集。
    實驗部分透過模擬資料與一個真實資料的實驗結果顯示MOMLGFM與TMOGFM是有效的。MOMLGFM的優點在能同時挖掘柏拉圖集合(隸屬函數集合)與多階層模糊關聯規則。TMOGFM的優點是不僅能挖掘隸屬函數集合,並且能挖掘各群組內的代表解並用於探勘多階層模糊關聯規則與利潤模糊商品集。
    Transactions in real-world applications usually consist of quantitative values. Some fuzzy data mining approaches have thus been proposed for deriving linguistic rules from this kind of transactions. Since membership functions may have a critical influence on final mining results, several genetic-fuzzy mining approaches have then been proposed as well for mining appropriate membership functions and fuzzy association rules at the same time. Most of them, however, focus on single-level concept and consider only one objective function. In view of this, this thesis proposes two approaches for mining the Pareto set (a set of non-dominated membership functions) and multi-level fuzzy association rules, namely a Multi-Objective Multi-Level Genetic-Fuzzy Mining Algorithm (MOMLGFM) and a Two-Stage Multi-Objective Fuzzy Mining Algorithm (TMOGFM).
    In the first algorithm (MOMLGFM), it first encodes the membership functions of each item class (category) into a chromosome according to the given taxonomy. Two objective functions are then considered. The first one is the knowledge amount mined out in different concept levels, and the second one is the suitability of membership functions. The fitness value of each individual is then evaluated by these two objective functions. After the MOGA process terminates, various sets of membership functions could be used for deriving multi-level fuzzy association rules according to decision makers’ preferences.
    However, the derived Pareto set by MOMLGFM may be not easy for users to choose an appropriate one for mining rules. In the second algorithm (TMOGFM), based on MOMLGFM, a two-stage multi-objective fuzzy mining algorithm is proposed for assisting decision makers to choose the proper solution. In the first stage, the MOMLGFM is used to derive a set of non-dominated membership functions (Pareto solutions). Then, in second stage, according to the designed rule-oriented or utility-oriented clustering attributes, the clustering technique is utilized to divide the Pareto solutions into groups and find representative solution of each group. The representative solutions of groups could be employed to mine fuzzy association rules or utility fuzzy itemsets according to the favorites of decision makers.
    Experimental results on simulation datasets and a real dataset also show the effectiveness of MOMLGFM and TMOGFM. The advantage of MOMLGFM is that it can derive Pareto set (a set of membership functions) and multi-level fuzzy association rules, simultaneously. The advantage of TMOGFM is that it can not only mine the Pareto set, but also use the representative solutions of groups to acquire multi-level fuzzy association rules and utility fuzzy itemsets.
    Appears in Collections:[資訊工程學系暨研究所] 學位論文

    Files in This Item:

    File SizeFormat
    index.html0KbHTML104View/Open

    All items in 機構典藏 are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library & TKU Library IR teams. Copyright ©   - Feedback