商業智慧實務 Practices of Business Intelligence

Presentation on theme: "商業智慧實務 Practices of Business Intelligence"— Presentation transcript:

1 商業智慧實務 Practices of Business Intelligence
Tamkang University 商業智慧實務 Practices of Business Intelligence 商業智慧的資料探勘 (Data Mining for Business Intelligence) 1022BI06 MI4 Wed, 9,10 (16:10-18:00) (B113) Min-Yuh Day 戴敏育 Assistant Professor 專任助理教授 Dept. of Information Management, Tamkang University 淡江大學 資訊管理學系 tku.edu.tw/myday/

2 課程大綱 (Syllabus) 週次 (Week) 日期 (Date) 內容 (Subject/Topics) 1 103/02/19 商業智慧導論 (Introduction to Business Intelligence) 2 103/02/26 管理決策支援系統與商業智慧 (Management Decision Support System and Business Intelligence) 3 103/03/05 企業績效管理 (Business Performance Management) 4 103/03/12 資料倉儲 (Data Warehousing) 5 103/03/19 商業智慧的資料探勘 (Data Mining for Business Intelligence) 6 103/03/26 商業智慧的資料探勘 (Data Mining for Business Intelligence) 7 103/04/02 教學行政觀摩日 (Off-campus study) 8 103/04/09 資料科學與巨量資料分析 (Data Science and Big Data Analytics)

3 課程大綱 (Syllabus) 週次 日期 內容（Subject/Topics） 9 103/04/16 期中報告 (Midterm Project Presentation) /04/23 期中考試週 (Midterm Exam) /04/30 文字探勘與網路探勘 (Text and Web Mining) /05/07 意見探勘與情感分析 (Opinion Mining and Sentiment Analysis) /05/14 社會網路分析 (Social Network Analysis) /05/21 期末報告 (Final Project Presentation) /05/28 畢業考試週 (Final Exam)

4 A Taxonomy for Data Mining Tasks
Source: Turban et al. (2011), Decision Support and Business Intelligence Systems

5 Market Basket Analysis
Source: Han & Kamber (2006)

6 Association Rule Mining
Apriori Algorithm Source: Turban et al. (2011), Decision Support and Business Intelligence Systems

7 Basic Concepts: Frequent Patterns and Association Rules
Itemset X = {x1, …, xk} Find all the rules X  Y with minimum support and confidence support, s, probability that a transaction contains X  Y confidence, c, conditional probability that a transaction having X also contains Y Transaction-id Items bought 10 A, B, D 20 A, C, D 30 A, D, E 40 B, E, F 50 B, C, D, E, F Customer buys diaper buys both buys beer Let supmin = 50%, confmin = 50% Freq. Pat.: {A:3, B:3, D:4, E:3, AD:3} Association rules: A  D (60%, 100%) D  A (60%, 75%) A  D (support = 3/5 = 60%, confidence = 3/3 =100%) D  A (support = 3/5 = 60%, confidence = 3/4 = 75%) Source: Han & Kamber (2006)

8 Market basket analysis
Example Which groups or sets of items are customers likely to purchase on a given trip to the store? Association Rule Computer  antivirus_software [support = 2%; confidence = 60%] A support of 2% means that 2% of all the transactions under analysis show that computer and antivirus software are purchased together. A confidence of 60% means that 60% of the customers who purchased a computer also bought the software. Source: Han & Kamber (2006)

9 Association rules Association rules are considered interesting if they satisfy both a minimum support threshold and a minimum confidence threshold. Source: Han & Kamber (2006)

10 Frequent Itemsets, Closed Itemsets, and Association Rules
Support (A B) = P(A  B) Confidence (A B) = P(B|A) Source: Han & Kamber (2006)

11 Support (A B) = P(A  B) Confidence (A B) = P(B|A)
The notation P(A  B) indicates the probability that a transaction contains the union of set A and set B (i.e., it contains every item in A and in B). This should not be confused with P(A or B), which indicates the probability that a transaction contains either A or B. Source: Han & Kamber (2006)

12 itemset K-itemset Example:
A set of items is referred to as an itemset. K-itemset An itemset that contains k items is a k-itemset. Example: The set {computer, antivirus software} is a 2-itemset. Source: Han & Kamber (2006)

13 The set of frequent k-itemsets is commonly denoted by LK
If the relative support of an itemset I satisfies a prespecified minimum support threshold, then I is a frequent itemset. i.e., the absolute support of I satisfies the corresponding minimum support count threshold The set of frequent k-itemsets is commonly denoted by LK Source: Han & Kamber (2006)

14 the confidence of rule A B can be easily derived from the support counts of A and A  B.
once the support counts of A, B, and A  B are found, it is straightforward to derive the corresponding association rules AB and BA and check whether they are strong. Thus the problem of mining association rules can be reduced to that of mining frequent itemsets. Source: Han & Kamber (2006)

15 Transactional data for an AllElectronics branch
Source: Han & Kamber (2006)

16 Example: Apriori Let’s look at a concrete example, based on the AllElectronics transaction database, D. There are nine transactions in this database, that is, |D| = 9. Apriori algorithm for finding frequent itemsets in D Source: Han & Kamber (2006)

17 Example: Apriori Algorithm Generation of candidate itemsets and frequent itemsets, where the minimum support count is 2. Source: Han & Kamber (2006)

18 Example: Apriori Algorithm
C1  L1 Source: Han & Kamber (2006)

19 Example: Apriori Algorithm
C2  L2 Source: Han & Kamber (2006)

20 Example: Apriori Algorithm
C3  L3 Source: Han & Kamber (2006)

21 The Apriori algorithm for discovering frequent itemsets for mining Boolean association rules.
Source: Han & Kamber (2006)

22 Generating Association Rules from Frequent Itemsets
Source: Han & Kamber (2006)

23 Example: Generating association rules
frequent itemset l = {I1, I2, I5} If the minimum confidence threshold is, say, 70%, then only the second, third, and last rules above are output, because these are the only ones generated that are strong. Source: Han & Kamber (2006)

24 關聯分析衡量的機率統計值— Support & Confidence
A B C D E B C E Rule A  D C  A A  C B & C  D Support 2/5 1/5 Confidence 2/3 2/4 1/3 Source: SAS Enterprise Miner Course Notes, 2014, SAS

25 Support & Confidence 高的關聯規則就一定是有用的規則？
Checking Account No Yes 4,000 500 3500 No Saving Account 6,000 1000 5000 Yes 10,000 Support(SVG  CK) = 50%=5,000/10,000 Confidence(SVG  CK) = 83%=5,000/6,000 Expected Confidence(SVG  CK) = 85%=8,500/10,000 Lift (SVG  CK) = Confidence/Expected Confidence = 0.83/0.85 < 1 Source: SAS Enterprise Miner Course Notes, 2014, SAS

26 關聯分析衡量的機率統計值— Lift增益值
信心水準最高的就是最好的規則? 「如果 Saving account 則 Checking account」這個規則的發生機率 比單獨計算Checking account的發生機率還低。 增益值(Lift)： 一條規則在預測結果時能比 隨機發生的機會好多少。 Lift (SVG  CK) = Confidence/Expected Confidence = 0.83/0.85 < 1 Source: SAS Enterprise Miner Course Notes, 2014, SAS

27 Support (AB) Confidence (AB) Expected Confidence (AB) Lift (AB)

28 Support (A B) = P(A  B) Confidence (A B) = P(B|A)
A與B 共同出現次數／總交易次數 Count(A&B)/Count(Total) Confidence (A B) = P(B|A) Conf (A  B) = Supp (A  B)/ Supp (A) A與B 共同出現次數／A出現的次數 Count(A&B)/Count(A) Expected Confidence (AB) = Support(B) Count(B) Lift (A  B) = Confidence (AB) / Expected Confidence (AB) Lift (A  B) = Supp (A  B) / (Supp (A) x Supp (B)) Lift (Correlation) Lift (AB) = Confidence (AB) / Support(B)

29 Lift (AB) Lift (AB) = Confidence (AB) / Expected Confidence (AB) = Confidence (AB) / Support(B) = (Supp (A&B) / Supp (A)) / Supp(B) = Supp (A&B) / Supp (A) x Supp (B) Lift 增益值 (提升值) Lift (AB) = 2 表示 AB 這條規則的增益值為 2， 代表已知在買A的前題下又買B的機率， 比直接買B 的機率提升 (增益)了2倍。

30 「買芭比娃娃就會買糖果」 你的行銷策略如何？
把兩項商品擺在一起 特意把兩項商品擺在相距較遠的地方 將糖果和芭比娃娃組合起來一起賣 糖果+芭比娃娃+銷售較差的商品一起組合銷售 定價策略：提供一個單價，降低另一個商品價格 廣告策略：芭比娃娃和糖果不需要同時廣告活動 產品設計：設計芭比娃娃形狀的糖果 提供芭比娃娃的配件，提升銷售 Source: SAS Enterprise Miner Course Notes, 2014, SAS

31 我的資料適合進行 購物籃分析嗎？ D A B Source: SAS Enterprise Miner Course Notes, 2014, SAS

32 Web Site Usage Associations
個案分析與實作二 (SAS EM 關連分析)： Case Study 2 (Association Analysis using SAS EM) Web Site Usage Associations

33 網站使用行為關聯分析

34 案例情境 ABC音樂廣播電台為了服務更多聽眾，設置了電台網站，讓更多的線上聽眾也可以透過網站服務以隨時掌握電台的各個節目資訊，網站提供了流行音樂趨勢(music streams)、音樂下載(podcasts)、新聞訊息(news streams)、線上收聽(live Web )以及歷史節目收聽(archives)等服務功能頁面。 分析人員想要藉由關聯分析以進一步了解線上聽眾的使用行為，做為網站服務功能更新的依據。 分析樣本為撈取近兩個月約150萬筆的客戶交易資料。 Source: SAS Enterprise Miner Course Notes, 2014, SAS

35 資料欄位說明 資料集名稱： webstation.sas7bdat ARCHIVE 廣播節目回顧 EXTREF 好站相連
LIVESTREAM 熱門節目收聽 MUSICSTREAM 流行音樂區 NEWS 最新消息 PODCAST 音樂下載 SIMULCAST 同步收聽 WEBSITE 首頁 Source: SAS Enterprise Miner Course Notes, 2014, SAS

36 網站使用行為關聯分析實機演練 分析目的 依據使用者網站交易資料，利用關聯分析演算方法產生網站使用行為關聯規則。 演練重點:
• 產生關聯分析資料集 • 進行關聯分析 • 關聯分析結果解釋 Source: SAS Enterprise Miner Course Notes, 2014, SAS

37 SAS Enterprise Miner (SAS EM) Case Study
Step 1. 新增專案 (New Project) Step 2. 新增資料館 (New / Library) Step 3. 建立資料來源 (Create Data Source) Step 4. 建立流程圖 (Create Diagram) SAS EM SEMMA 建模流程

38 Download EM_Data.zip (SAS EM Datasets)

39 Upzip EM_Data.zip to C:\DATA\EM_Data

40 Upzip EM_Data.zip to C:\DATA\EM_Data

41 VMware Horizon View Client softcloud.tku.edu.tw SAS Enterprise Miner

42 SAS Enterprise Guide (SAS EG)

43 SAS EG New Project

44 SAS EG Open Data

45 SAS EG Open webstation.sas7bdat

46 webstation.sas7bdat

47 webstation.sas7bdat

48 SAS Enterprise Miner 12.1 (SAS EM)

49 SAS EM 資料匯入4步驟 Step 1. 新增專案 (New Project)
Step 2. 新增資料館 (New / Library) Step 3. 建立資料來源 (Create Data Source) Step 4. 建立流程圖 (Create Diagram)

50 Step 1. 新增專案 (New Project)

51 Step 1. 新增專案 (New Project)

52 Step 1. 新增專案 (New Project)

53 SAS Enterprise Miner (EM_Project2)

54 Step 2. 新增資料館 (New / Library)

55 Step 2. 新增資料館 (New / Library)

56 Step 2. 新增資料館 (New / Library)

57 Step 2. 新增資料館 (New / Library)

58 Step 2. 新增資料館 (New / Library)

59 Step 3. 建立資料來源 (Create Data Source)

60 Step 3. 建立資料來源 (Create Data Source)

61 Step 3. 建立資料來源 (Create Data Source)

62 Step 3. 建立資料來源 (Create Data Source)

63 Step 3. 建立資料來源 (Create Data Source)

64 Step 3. 建立資料來源 (Create Data Source)
DatabaseName.TableName LibraryName.TableName EM_LIB.WEBSTATION

65 Step 3. 建立資料來源 (Create Data Source)

66 Step 3. 建立資料來源 (Create Data Source)

67 Step 3. 建立資料來源 (Create Data Source)

68 Step 3. 建立資料來源 (Create Data Source)

69 Step 3. 建立資料來源 (Create Data Source)

70 Step 3. 建立資料來源 (Create Data Source)

71 Step 3. 建立資料來源 (Create Data Source)

72 Step 3. 建立資料來源 (Create Data Source)
Data Source Attribute Role: Transaction

73 Step 3. 建立資料來源 (Create Data Source)

74 Step 3. 建立資料來源 (Create Data Source)

75 Step 4. 建立流程圖 (Create Diagram)

76 Step 4. 建立流程圖 (Create Diagram)

77 Step 4. 建立流程圖 (Create Diagram)

78 SAS Enterprise Miner (SAS EM) Case Study
Step 1. 新增專案 (New Project) Step 2. 新增資料館 (New / Library) Step 3. 建立資料來源 (Create Data Source) Step 4. 建立流程圖 (Create Diagram) SAS EM SEMMA 建模流程

79 案例情境模型流程

80 樣本資料匯入 (Sample)

81 EM_Lib.Webstation

82 樣本資料匯入 (Sample) Edit Variable

83 樣本資料匯入 (Sample) Edit Variable - Explore …

84 樣本資料匯入 (Sample) Edit Variable - Explore …

85 Explore - Association

86 關聯分析 (Association Analysis)

87 關聯分析 (Association Analysis)

88 關聯分析 (Association Analysis)

89 關聯分析 (Association Analysis)

90 關聯分析 (Association Analysis)

91 關聯分析 (Association Analysis)

92 關聯分析 (Association Analysis)

93 關聯分析 (Association Analysis) Support : 1% (Minimum Support = 1%)

94 關聯分析 (Association Analysis)

95 關聯分析 (Association Analysis)

96 關聯分析 (Association Analysis) 檢視/規則/規則表格 (Rules Table)

97 關聯分析 (Association Analysis) Association Rules - 規則表格 (Rules Table)

98 關聯分析 (Association Analysis) Association Rules - 規則表格 (Rules Table)

99 關聯分析 (Association Analysis) 檢視/規則/連結圖形 (Link Graph)

100 關聯分析 (Association Analysis) 連結圖形 (Link Graph)

101 關聯分析 (Association Analysis) Maximum Number of Items: 3000000

102 關聯分析 (Association Analysis)

103 關聯分析 (Association Analysis) Association Rules - 規則表格 (Rules Table)

104 關聯分析 (Association Analysis) 連結圖形 (Link Graph)

105 References Efraim Turban, Ramesh Sharda, Dursun Delen, Decision Support and Business Intelligence Systems, Ninth Edition, 2011, Pearson. Jiawei Han and Micheline Kamber, Data Mining: Concepts and Techniques, Second Edition, 2006, Elsevier Jim Georges, Jeff Thompson and Chip Wells, Applied Analytics Using SAS Enterprise Miner, SAS, 2010 SAS Enterprise Miner Course Notes, 2014, SAS SAS Enterprise Miner Training Course, 2014, SAS SAS Enterprise Guide Training Course, 2014, SAS