DATA MINING USING CLUSTERING METHODS
TO SUPPORT DECISION MAKING FOR SALES
SERVICE LEVEL AT CABLE BROADCAST SUBSCRIPTIONS
RESEARCH
BOBBY SANJAYA (0932401824)
Graduate Program in Computer Studies Master of Information Technology
Bina Nusantara University Jakarta
DATA MINING USING CLUSTERING METHODS
TO SUPPORT DECISION MAKING FOR SALES
SERVICE LEVEL AT CABLE BROADCAST SUBSCRIPTIONS
RESEARCH
BOBBY SANJAYA (0932401824)
A Thesis Submitted to the Faculty In Partial Fulfillment of the Requirements For Master Degree in Information Technology
DATA MINING USING CLUSTERING METHODS
TO SUPPORT DECISION MAKING FOR SALES
SERVICE LEVEL AT CABLE BROADCAST SUBSCRIPTIONS
RESEARCH
BOBBY SANJAYA (0932401824)
1st Supervisor : 2nd Supervisor:
Bambang Heru Iswanto, S.Si., M.Si., Dr.rer.nat
Agus Widodo, B.Sc., MT.
PERNYATAAN
STATEMENT
Dengan ini saya,
Nama : Bobby Sanjaya NIM : 0932401824
Judul tesis : Data Mining Using Clustering Methods To Support Decision Making For Sales Service Level At Cable Broadcast Subscription
Memberikan kepada Universitas Bina Nusantara hak non-eksklusif untuk menyimpan, memperbanyak, dan menyebarluaskan tesis karya saya secara keseluruhan atau hanya sebagian atau hanya ringkasannya saja, dalam bentuk format tercetak dan atau elektronik.
Menyatakan bahwa saya akan mempertahankan hak exclusive saya untuk menggunakan seluruh atau sebagian isi tesis sayaguna pengembangan karya di masa depan, misalnya bentuk artikel, buku, perangkatlunak, ataupun sistem informasi.
Hereby grant to my school, Bina Nusantara University , the non-exclusive right to archive, reproduce, and distribute my thesis, in whole or in part , whether in the form of printed and electronic formats.
I acknowledge that I retain exclusive rights of my thesis by using all or part of it in the future work or outputs, such as article, book, software, and information system.
Jakarta, 16Agustus 2011
HALAMAN PERNYATAAN
STATEMENT PAGE
Saya, nama Bobby Sanjaya, NIM 0932401824 menyatakan dengan sebenar-benarnya bahwa tesis saya berjudul “Data Mining dengan Metoda-metoda Clustering Untuk Mendukung Pengambilan Keputusan Pada Service Level Sales Dari Perusahaan Kabel Tayang” adalah merupakan gagasan dan hasil penelitian/proyek saya sendiri dengan bimbingan Dosen Pembimbing.
Saya juga menyatakan dengan sebenarnya bahwa isi tesis ini tidak merupakan jiplakan dan bukan pula dari karya orang lain, kecuali kutipan dari literature dan atau hasil wawancara tertulis yang saya acu dan telah saya sebutkan di Daftar Acuan dan Daftar Pustaka.
Demikian pernyataan ini saya buat dengan sebenarnya dan saya bersedia menerima sanksi apabila ternyata pernyataan saya ini tidak benar.
I, NameBobby Sanjaya, Student ID 0932401824 truly acknowledge that my thesis with
title “Data Mining Using Clustering Methods To Support Decision Making For Sales Service Level At Cable Broadcast Subscription” is my concept and project result with guidance from supervisor.
I, also truly acknowledge that content of this thesis are not copyed and not from another people work, except my citation from literature or written interview result and already write in reference list and bibliography list.
That’s my acknowledge were truly made and if in reality this acknowledge weren’t true, I willing sanction.
Jakarta, 16 Agustus 2011 Yang menyatakan
Bobby Sanjaya 0932401824
AKNOWLEDGEMENTS
I would like to offer my gratitude to God Almighty for His guidance during the making of this thesis. I also offer my deepest gratitude for all of those who has helped me:
1. My thesis counselors, Bambang Heru Iswanto, S.Si., M.Si., Dr.rer.nat and Agus Widodo, B.Sc., MT., thank you for the guidance for the subject of interest.
2. Every lecturer I met during my master study that has given me the idea for what should I follow.
3. My sister whom has given many encourage for me to go through this trial.
4. My friends from college and workplaces, every support you have given to me really means very much.
5. To every family and relations that has given me their advices and support.
I hope that this thesis would be useful for future study and therefore I offer my humble apology for anything lacks from this writings.
Jakarta, August2011
TABLE OF CONTENT
Title ……… i
Statement Title …...………. ii
Supervisor Approval ………. iii
Statement ………. iv
Statement Page ………. v
Acknowledgements ………. vi
Abstract ………. vii
Table of Content ………. viii
List of Figures …………..……… x List of Tables ………..……… xi Chapter 1. Introduction ……… 1 1. 1. Background ……… 1 1. 2. Statement of Problems ………..……….. 3 1. 3. Goal of Study ……… 3 1. 4. Benefit of Study ……… 4
Chapter 2. Literature Review ……… 5
2. 1. Data Mining ………..……….. 5
2. 1. 1.Clustering Methods ………….………..…. 7
2. 1. 1. 1. Self-Organizing Maps ……… 9
2. 1. 1. 2. K-Means ………... 14
2. 1. 1. 3. Expectation-Maximization Mixture Model ………. 17
2. 2. Principal Component Analysis ………. 20
2. 3.Normalization ………. 22
2. 5. Service Level in Sales ……… 24 2. 5. 1. Direct Sales ……… 26 2. 5. 2. Market Place ……… 27 2. 5. 3. Multi-Level Marketing ……… 27 2. 8. 4. Strategies ……… 28 2. 4. Study Motivation ……….. 32 Chapter 3. Methodology ……… 39 3. 1. Conceptual Study ……… 39 3. 2. Study Purpose ……… 40 3. 3. Preprocessing ……… 41 3. 3. 1. Feature Selection ……… 42 3. 3. 2. Normalization ……… 42
3. 3. 3. Principal Component Analysis ……… 43
3. 4. Clustering Analysis ……….. 43
3. 5. Validation and Evaluation ……… 47
3. 6. Data Mining Tools ………. 47
Chapter 4. Results and Analysis ……… 49
4. 1. Preprocessing ……… 49
4. 2. Principal Component Analysis ……… 53
4. 3. Cluster Analysis ……… 53
4. 4. Direct Sales Clustering ……… 55
4. 5. Modern Stores Clustering ……… 63
4. 6. Multi-Level Marketing (MLM) Clustering ……… 70
4. 7. Result Analysis ……… 76
Chapter 5. Conclusion ………. 79
References ………. 81
LIST OF FIGURES
Figure 2. 1. Example of How SOM Works in Detecting Clusters by Defining (Upper) and
Locating the Intended Cluster (Below) ……… 10
Figure 2. 2. SOM training process ………. 12
Figure 2. 3. Clustering results using 10 clusters (shown by projection into 2d space) using the sample training image. K-means clustering without applying kernel transformation (left) is compared with kernel k-means clustering (right) ………. 14
Figure 2. 4. K-Means distant and close nodes ………. 16
Figure 2. 5. GMM fitting examples from EM estimates ……… 18
Figure 2. 6. Process Thinking and Learning Hierarchy ………. 32
Figure 2. 7. The Breadth and Depth of CRM Process Organization ………. 32
Figure 2. 8. The Centers of Five Clusters Compared on the Same Graph. This Simple Visualization Technique (called Parallel Coordinates) Helps Identify Interesting Clusters 34 Figure 2. 9.SOM learning process ……… 36
Figure 2. 10.Performance evalutation of previous cluster research method ……… 37
Figure 3. 1. Conceptual study of research ……… 39
Figure 3. 2. Research Process ………. 40
Figure 3. 3. Flowchart of clustering process for this study ………. 46
Figure 4. 1. Synapse’s SOM Training sessions ………. 54
Figure 4. 2. SOM cluster mapping ………. 58
Figure 4. 3. SOM cluster mapping for Modern Stores ……… 65
LIST OF TABLES
Table 2. 1. Generated Meta-Thematic Categories ………. 30
Table 4. 1a. Result of data collection, before initial data preprocessing ………. 50
Table 4. 1b. Result of data collection, after initial data preprocessing ………. 51
Table 4. 2. Result after normalization ………. 52
Table 4. 3. Results from univariate continuous statistics ………. 52
Table 4. 4. PCA analysis ………. 53
Table 4. 5. Summary of SOM cluster training ………. 55
Table 4. 6. SOM on Direct Sales ………. 56
Table 4. 7. Direct Sales cluster summaries ………. 57
Table 4. 8. K-Means for Direct Sales ………. 59
Table 4. 9. Summary of K-Means for Direct Sales ………. 60
Table 4. 10. Contingency Chi-Square analysis on SOM & K-Means clusters . 60 Table 4. 11. EM for Direct Sales ………. 61
Table 4. 12. EM Summaries for Direct Sales ………. 61
Table 4. 13. Contingency Chi-Square analysis on SOM & EM clusters ………. 62
Table 4. 14. Contingency Chi-Square analysis on K-Means & EM clusters ………. 62
Table 4. 15. SOM on Modern Stores ………. 63
Table 4. 16. Modern Stores SOM cluster summaries ………. 64
Table 4. 17. K-Means for Modern Stores ………. 66
Table 4. 18. Summary of K-Means for Modern Stores ………. 66
Table 4. 19. Modern Stores Chi-Square analysis on SOM & K-Means clusters . 67 Table 4. 20. EM for Modern Stores ………. 68
Table 4. 21. EM Summaries for Modern Stores ………. 68
Table 4. 24. SOM on MLM ……… 70
Table 4. 25. SOM cluster summaries for MLM ……… 70
Table 4. 26. K-Means for MLM ……… 72
Table 4. 27. Summary of K-Means for MLM ……… 72
Table 4. 28. MLM’s Chi-Square analysis on SOM & K-Means clusters ……… 73
Table 4. 29. EM for MLM ……… 74
Table 4. 30. EM Summaries for MLM ……… 74
Table 4. 31. MLM’s Chi-Square analysis on SOM & EM clusters ……… 75
