臺灣博碩士論文加值系統

近來資訊科技的蓬勃發展，使得企業能夠輕易的獲得多元化的資料，在這大量的資料當中，潛藏了許多對企業有用的資訊，可以提供給企業對於經營策略的制定作為參考的依據，而資料探勘(data mining)這種資料分析方法可以找出潛藏在資料中的資訊。分類分析(classification)中的決策樹(decision trees)是資料探勘技術中常被使用的一種資料探勘方法，依據已知的類別建立預測性模型，以解決類別之特性，企業可以根據這個預測性模型來找出產品屬性等屬性間的關聯性，再利用這些屬性間的關聯性來訂定產品行銷等企業經營管理策略。
以往的資料探勘技術中大多是對資料庫直接進行探勘，實際上，由於產品利潤的差異，其在產品行銷組合的重要性也會有所不同，若使用原始的分類分析探勘方法，很容易會忽略了一些獲利率高、但交易次數不頻繁的產品。因此藉由加權來改變其在總交易資料數量中出現次數的計算值，發掘考量加權因素的產品組合型樣。每一筆交易資料除了記錄有消費者曾經購買過的產品項目，也記錄著其購買時的次序性，顯示出消費者是依序本身的需求及產品的特性來有次序地購買產品項目，因此利用分類技術來發掘具有購買次序的產品組合型樣。這些探勘結果對於企業在擬訂產品之行銷組合的策略時，可以提供非常有用的參考資訊。
在傳統建構決策樹的過程中，由於大量交易資料的讀取造成資源上的浪費，為了提升探勘的執行效率，本研究使用布林運算為基礎以發掘產品項目之行銷組合型樣。將原先的交易資料庫進行編碼轉換，把資料轉換成bit的格式，使用“0”與“1”表示資料的交易狀況，運算的時候可以使用轉換後的資料進行布林運算，可提升資料在探勘時的執行效率。

With flourishing development of information science and technology, the establishment can get easily more and more original data and materials that are including lots of potential and useful information that can offer establishment the formulation of management strategies as a reference. Data mining is an analytic skill that can analyze and find out the potential information among the original data and materials. Under data mining, the way used most frequently for prospecting original materials is decision tree of classification. According this divinable model, business can find out the relation between all attributes such as customers, products and regions etc, so that they can formulate some business management strategies like product-marketing base on these relations.
Formerly, most of mining technology were prospected the database directly. However, because of the profits of product, the improve of the product marketing will have different. If use the primitive classification. So, some of high profit and few trading frequency products, shall be ignored easily. We will mine marketing mix patterns of product items with weighted that product items will adjust the number of items in the trading materials. Additionally, we will mine marketing mix patterns of product items from the trading materials record of the sequence. These results of mining are very useful information for enterprises to draft the strategy.
Traditionally, during the process of building the decision trees, the reading of a large amount of trading materials causes the waste of resources. In order to improve the executive efficiency, we will mine marketing mix patterns of product items based on Boolean algorithm. Changing the trading materials into forms of bit “0” and “1” by operate with Boolean algorithm that usually needs to scan the trading materials in order to improve the efficiency of prospecting.
Keywords: data mining, classification, weighted, sequence, Boolean algorithm, marketing mix

第一章 緒論.........................................................1
1.1 研究動機................................................1
1.2 研究目的................................................2
1.3 研究範圍與限制...........................................3
1.4 研究流程................................................4
1.5 論文架構................................................5
第二章 文獻探討..................................................6
2.1資料庫知識發掘............................................6
2.2分類分析.................................................9
2.2.1 決策樹意義........................................ 9
2.2.2 決策樹的定義.......................................9
2.3資料探勘相關技術.........................................11
2.3.1 ID3演算法........................................11
2.3.2 布林演算法.......................................15
2.3.3 建構決策樹之相關演算法............................16
第三章 發掘產品項目最適性之行銷組合...............................17
3.1利用ID3演算法發掘產品項目最適性之行銷組合..................17
3.1.1探勘方法.............................................17
3.1.2實例說明...........................................19
3.2考量加權因素發掘產品項目最適性之行銷組合...................22
3.2.1探勘方法.............................................22
3.2.2實例說明...........................................24
3.3利用分類技術發掘具有購買次序之產品項目.....................28
3.3.1探勘方法.............................................28
3.3.2實例說明...........................................31
第四章 以布林運算為基礎發掘產品項目最適性之行銷組合.................37
4.1以布林運算為基礎的ID3演算法..............................37
4.1.1建構方法...........................................37
4.1.2實例說明...........................................40
4.2以布林運算為基礎建構考量加權因素之ID3決策樹................45
4.2.1建構方法...........................................45
4.2.2實例說明...........................................46
4.3以布林運算為基礎建構具有購買次序之ID3決策樹................54
4.3.1建構方法...........................................54
4.3.2實例說明...........................................56
第五章 效能分析................................................68
5.1實驗設計...............................................68
5.2實驗過程...............................................70
5.3實驗結果...............................................79
5.3.1 ID3演算法的探勘效..................................79
5.3.2考量加權因素之ID3演算法探勘效能.......................80
5.3.3具有購買次序之ID3演算法探勘效能.......................81
5.3.4不同產品項目之ID3演算法的探勘效能.....................82
5.3.5不同產品項目之考量加權因素ID3演算法探勘效能.............83
5.3.6不同產品項目之具有購買次序ID3演算法探勘效能.............84
第六章 結論與未來發展...........................................85
6.1研究結論................................................85
6.2未來研究方向............................................86
參考文獻

一、中文部分
1.魏志平、董和昇，電子商豫理論與實務(2版)，華泰書局，pp. 167-205，2002。
2.陳垂呈，以有效率的布林演算法來擷取關聯規則，2002數位生活與網際網路科技研討會，台南，成功大學，六月，2002。
3.顏秀珍，邱鼎穎，李御璽，從大型資料庫中挖掘加權的關聯規則，全國計算機會議，2001。
二、英文部分
4.P. Smyth, S. G. Piatetsky and U. Fayyad, “From Data Mining to Knowledge Discovery in Database,” Cambridge, AAA/MIT Press, 1996.
5.M. S. Chen, J. Han, and P. S Yu, “Data Mining: An Overview from a Database Perspective,” IEEE Trans. on Knowledge and Data Engineering, Vol. 8, No. 6, pp. 866-883, 1996
6.M. J. A. Berry and G. Linoff, Data Mining Techniques for Marketing Sales and Customer Support, John Wiley & Sons, 1997.
7.L. Kaufman, and P. J. Rousseeuw, Finding Groups in Data: an Introduction to Cluster Analysis, John Wiley & Sons, 1990.
8.K. Alsabti, S. Ranka and V. Singh, “An Efficient K-Means Clustering Algorithm,” PPS/SPDP Workshop on High Performance Data Mining, 1997.
9.R. C. Dubes and A. K. Jain, Algorithms for Clustering Data, Prentice Hall, 1988.
10.R. T. Ng and J. Han, “Efficient and Effective Clustering Methods for Spatial Data Mining,” Proc. of the 20th International Conference on Very Large Data Bases, pp. 144-155, 1994.
11.J. R. Quinlan, “Introduction of Decision Trees,”　Machine Learning , Vol. 1, pp. 81-106, 1986.
12.S. Y. Wur, and Y. Leu, “An Effective Boolean Algorithm for Mining Association Rules in Large Databases,” DASFAA, 1999.
13.J. Yang, H. Wang, X. G.. Hu, and Z. H. Hu, “A New Classification Algorithm Based on Rough Set and Entropy,” International Conference on Machine Learning and Cybernetics, 2003.
14.Z. Meng and Z. Cai, “An Improved Algorithm of Decision Tree Based on Attribute Entropy,” Proc. of the 5th World Congress on Intelligent Control and Automation, pp. 4268-4272, 2004.
15.H. Ding, X. K. Wang, “Research on Algorithm of Decision Tree Induction,” Proc. of the First Int. Conference On Meachine Learn and Cybernetics, 2002.
16.Q. Yang, J. Yin, Charles X. Ling and Tielin Chen, “Postprocessing Decision Trees to Extract Actionable Knowledge,” Proc. of the Third IEEE Int. Con. On Data Mining, 2003.
17.K. W. Lam and V. C. S. Lee, “Building Decision Trees Using Functional Dependencies,” Pro. of the Int. Con. On Information Technology: Coding and Computing, 2004.
18.C. L. Sabharwal, K. R. Hacke and D. C. St. Clair, “Formation of Clusters and Resolution of Orkinal Attributes in ID3 Classification Trees,” Proc. of ACM/SIGAPP Symposium on Applied computing: Technological Challenges of the 1990’s, pp. 590-597, 1992.
19.C. H. Cai, W. C. Fu, C. H. Cheng and W. W. Kwong, “Mining Association Rules with Weighted Items,” International Database Engineering and Applications Symposium (IDEAS), pp. 68-77, 1998.
20.E. Rich and K. Knight, Learning in Neural Network, 2ndEd., McGraw-Hill, New York, 1991.
21.K. C. Laudon and J. P. Laudon, Management Information Systems, Sixth Edition, Upper Saddle River, NJ: Prentice Hall, 2000.
22.P. Clark and T. Niblett, “The CN2 Induction Algorithm,” Machine Learning, vol. 3, pp.261-283, 1989.
23.R. Agrawal and R. Srikant, “Mining Sequential Patterns,” Proceedings of the International Conference on Data Engineering(ICDE), 1995.
24.R. Skirkant and R. Agrawal, “Mining Sequential Patterns,” Generalizations and Performance Improvements,” Proceedings of the Fifth International Conference on Extending Database Technology (EDBT), 1996.
25.S. C. Hui and G. Jha, “Data Mining for Customer Service Support,” Information and Management, vol. 38, pp. 1-13, 2000.