模糊關聯規則於不明確資料問卷上之探勘研究

Study on the Mining Fuzzy Association Rules on Questionnaire Contains Uncertain Data

林欣弘(Hsin-Hung Lin)

指導教授 ： 楊達立

國立虎尾科技大學/管理學院/資訊管理研究所/碩士(2010年)

摘要

資料探勘技術近年來已被大量的使用於各領域中，大多認為資料中隱
藏了可能存在的知識，來增加我們對於事件的更深了解。資料探勘主要為對目標資料進行挖掘出有特徵樣式之程序，來解釋現存行為關係或預測未來的結果，其中關聯規則為對目標資料間資料關聯性的研究。現今有許多關聯性資料法則的研究，但大多數定義在所得資料為精確且確定性的研究上，然而在現實中此條件是不符合現況的，因為常常由於人們的種種疏失或是記錄上的瑕疵，會導致資料取得的不明確，或是問卷常為了研究方便性而採用序數值答題，而侷限了受測者的答題思維。本研究延續Weng[18]學者的模式下，提出一個結合資料探勘、模糊集合(Fuzzy sets)等技術的探勘關聯方式，從不明確的資料中探勘關聯規則，讓重要關聯規則不再因資料性質而有所遺漏，導致重要資訊不被採用。經由電腦模擬結果發現本研究所產生之關聯規則關聯度更佳且更多。最後使用英國汽車公司(Land Rove)所調查問卷資料利用Poly Analyst 軟體和本研究所提出方法做比較，證明本
研究所提出之演算法優於原先研究所獲得的結果。

關鍵字

資料探勘 ； 關聯規則 ； 模糊集合 ； 不明確性資料

並列摘要

In recent years, data mining has been widely used in various fields. Due to most of information may exist in the hidden knowledge and for promoting better understanding of the events. Data mining therefore is used to mining out the
characteristic style of the object data and explains the relationship between the existing behaviors or predicts future results. Today, although there are many rules of relevance information, but most are defined in the information available for the accurate and definitive research. However, this condition is incompatible with the current situation, because people often made various negligence or record defect. This will lead to information obtained is uncertain. Moreover, questionnaires often
adopt the numerical sequence answer for convenience, but limiting the thinking of the subjects. This thesis proposes a approach to mining association rules from the uncertain data by combining data mining, fuzzy set and other technology associated with mining methods by extending the study of Weng [18]. The important association rules will no longer are omitted due to the nature of the information by the proposed approach in this research. By computer simulation results showed that the proposed association rules have better and more relational grade in this study. Finally, the Land Rover (Discovery SUV) questionnaire data is used to compare the proposed method in this study by the use of Poly Analyst software. The results showed that the proposed algorithm is superior to the original research results obtained.

並列關鍵字

Data mining ； Association rules ； Fuzzy set ； Uncertain data

參考文獻

[1] Srikant R. andAgrawal R., Mining quantitative association rules in large relational tables, Proceedings of ACM SIGMOD , pp. 1-12, 1996.

[2] Agrawal R., Imielinski T. and Swami A., Mining association rules between sets of items in large databases, Proceedings of the SIGMOD'93, Washington, DC, pp. 207-216, 1993.

[3] Caskey K.R., ”A manufacturing problem solving environment combining evaluation, search, and generalisation methods,” Computers in Industry,Vol.44, No.2, pp.175-187, 2001.

[5] Kleissner C.,“Data mining for the enterprise,” In Proceedings of the Thirty-First Hawaii International Conference on, Volume 7,pp. 295-304, 1998.

[7] Han, J. and Micheline K. , “Data Mining : Concepts and Techniques, ”John Wiley & Son，2001.