以多維度倉儲系統支援校務研究決策分析用的因果規則發掘

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：39

、訪客IP：18.220.137.164

姓名

汪瑞勛(Jui-Hsun Wang) 查詢紙本館藏

畢業系所

軟體工程研究所

論文名稱

以多維度倉儲系統支援校務研究決策分析用的因果規則發掘
(Causal Rule Mining Based on Multi-Dimensional Structure for Institutional Research Decision-Support Task)

相關論文

★ 應用自組織映射圖網路及倒傳遞網路於探勘通信資料庫之潛在用戶	★ 基於社群網路特徵之企業電子郵件分類
★ 行動網路用戶時序行為分析	★ 社群網路中多階層影響力傳播探勘之研究
★ 以點對點技術為基礎之整合性資訊管理及分析系統	★ 在分散式雲端平台上對不同巨量天文應用之資料區域性適用策略研究
★ 應用資料倉儲技術探索點對點網路環境知識之研究	★ 從交易資料庫中以自我推導方式探勘具有多層次FP-tree
★ 建構儲存體容量被動遷徙政策於生命週期管理系統之研究	★ 應用服務探勘於發現複合服務之研究
★ 利用權重字尾樹中頻繁事件序改善入侵偵測系統	★ 有效率的處理在資料倉儲上連續的聚合查詢
★ 入侵偵測系統：使用以函數為基礎的系統呼叫序列	★ 有效率的在資料方體上進行多維度及多層次的關聯規則探勘
★ 在網路學習上的社群關聯及權重之課程建議	★ 在社群網路服務中找出不活躍的使用者

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

[檢視]

[下載]

本電子論文使用權限為同意立即開放。
已達開放權限電子全文僅授權使用者為學術研究之目的，進行個人非營利性質之檢索、閱讀、列印。
請遵守中華民國著作權法之相關規定，切勿任意重製、散佈、改作、轉貼、播送，以免觸法。

摘要(中)

校務研究目前在世界各國受到許多學校的重視，台灣近年來因為大專院校不
斷增設，衍生出許多問題，為了讓決策者可以從不同面向分析並做出最合適的決
策，政府機關與各高等教育機構亦開始極力推行校務研究，以更科學化論述事證
的方式面對環境與教育政策的改變。
校務研究涵蓋的範圍極為廣泛，包含學務、教務、人事以及資源等，目前校
務研究單位雖整合了學校各行政單位的資料，但仍有許多有價值的資料尚未被分
析和使用。本研究以學生為主要分析面向，每位學生背景不同，在學時參與的活
動、修習及獎懲等表現也不同，學生在校的歷程往往會對畢業後的發展趨勢有所
影響。藉由學生入學前的入學方式與高中類型，以及在學的社團參與、獎懲、工
讀等加上畢業的工作類型與薪資等分析學生發展的趨勢，對於學校評估學生學習
成效或是輔助校務決策有許多幫助。
為了由不同面向發掘資料的因果規則，本研究以多維度資料倉儲作為分析的
資料來源，將不同維度的屬性依照時間發生的順序形成序列，再透過循序樣式探
勘，探索學生在校狀況及畢業流向間有效的因果規則，以支援校務決策之重要分
析。

摘要(英)

Institutional Research has been a significant feature of higher education in other
countries for many years. In Taiwan, as the education policy and society environment
change rapidly in responding to pressures of globalization and competition, decision
makers need scientific methods to catch up with the times. Therefore, higher education
institutions are starting to implement institutional research these years.
The wide coverage of topic of institutional research includes academic affair,
student affair, personnel affair, general affair, etc. In IR data warehouse, integrated data
can be turned into meaningful information for decision making. This paper focuses on
finding causal rules from student portfolios including type of admissions, clubs,
rewards and punishments, part-time jobs and questionnaire survey of graduation to
analyze career development.
In order to support decision making, in this paper, data from multi-dimensional
data warehouse are seen to be employed as input data and an improved sequential
pattern mining method is proposed for discovering useful causal relations.

關鍵字(中)

★ 資料倉儲
★ 因果規則
★ 循序樣式探勘
★ 校務研究

關鍵字(英)

★ Data Warehousee
★ Causal Rule
★ Sequential Pattern Mining
★ Institutional Research

論文目次

摘要 ................................................................................................................................. i
Abstract ..........................................................................................................................ii
誌謝 .............................................................................................................................. iii
目錄 ............................................................................................................................... iv
一、序論 ............................................................................................................. 1
1-1 研究背景及動機 ............................................................................................ 1
1-2 研究目的 ........................................................................................................ 2
1-3 論文架構介紹 ................................................................................................ 4
二、文獻探討 .................................................................................................. 5
2-1 校務研究 (Institutional Research，IR) ........................................................ 5
2-2 多維度資料倉儲 (Multi-Dimensional Data-Warehouse) ............................ 8
2-3 資料方塊 (OLAP Cube) ............................................................................. 10
2-4 多維度資料模型 (Multidimensional Schema) ........................................... 14
2-5 關聯規則 (Association Rule) ..................................................................... 17
2-6 循序樣式探勘 (Sequential Pattern Mining) ............................................... 17
2-7 勝算比 (Odds Ratio) ................................................................................... 19
三、系統架構與流程 ................................................................................ 20
3-1 系統流程 ...................................................................................................... 20
3-2 前處理過程與建置 ...................................................................................... 21
3-3 因果規則查詢 .............................................................................................. 21
3-4 以不同面向進行因果規則探勘 .................................................................. 21
四、研究方法 ................................................................................................ 22
4-1 資料前處理–資料萃取 ................................................................................ 22
4-2 資料前處理–資料轉置 ................................................................................ 25
4-3 多維度因果規則計算 .................................................................................. 27
4-4 決策變數精細度轉換 .................................................................................. 28
v

4-5 循序樣式查詢 .............................................................................................. 29
4-6 以勝算率加強驗證因果規則 ...................................................................... 30
4-7 觀察面向一般化 .......................................................................................... 31
4-8 探索不同結果 .............................................................................................. 32
4-9 使用者查詢介面 .......................................................................................... 33
五、實作 ........................................................................................................... 35
5-1 實驗規格與環境 .......................................................................................... 35
5-2 資料建構階段實驗 ...................................................................................... 36
5-2-1 多維度資料建立的時間比較 .............................................................. 36
5-2-2 資料載入的時間比較 .......................................................................... 37
5-3 因果規則探勘比較 ...................................................................................... 38
5-3-1 相同因素比較-社團參與 .................................................................... 39
5-3-2 相同因素比較-工作情況 .................................................................... 41
5-3-3 不同因素比較-校內工讀 .................................................................... 42
5-3-4 不同因素比較-畢業後工作情況 ........................................................ 43
六、結論 ........................................................................................................... 45
七、參考文獻 ................................................................................................ 46

參考文獻

[1] Institutional Research, https://en.wikipedia.org/wiki/Institutional_research
[2] Assocaition for Institutional Resaerch, https://www.airweb.org/
[3] 彭森明, 政府如何協助大專院校推展校務研究 : 美國經驗, 2015
[4] 高等教育深耕計畫,
https://zh.wikipedia.org/wiki/%E9%AB%98%E7%AD%89%E6%95%99%E8%
82%B2%E6%B7%B1%E8%80%95%E8%A8%88%E7%95%AB
[5] Michael Yao-Ping Peng and Sophia Shi-Huei Ho, “Strategic Agenda-Setting of
Institutional Research in Taiwan′s Higher Education Institutions”, 2016
[6] TajimaYushi, “On Practical Institutional Research for Small-Scale Art College: A
Case Study of Takarazuka University of Art and Design Tokyo School of Media
Art”, 2017 6th IIAI International Congress on Advanced Applied Informatics
(IIAI-AAI), 2017
[7] Tong Wang, “Application of Data Warehouse and OLAP Technology in Students′
Teaching Evaluation”, 2018 IEEE/ACIS 17th International Conference on
Computer and Information Science (ICIS), 2018
[8] Yisong Huang, Jingjing Yin and Hani Samawi, “Methods improving the estimate
of diagnostic odds ratio”, Communications in Statistics - Simulation and
Computation, 2016
[9] Howard D.Richard, McLaughlin W.Gerald and Knight E.William, “The
Handbook of Institutional Research”, Jossey-Bass Publishers, 2012
[10] Taiwan Association for Institutional Research, http://www.tair.tw/en
US/about/mission
[11] 王麗雲, 透過校務研究進行自我評鑑與自我改進, 2014,
http://epaper.heeact.edu.tw/archive/2014/01/01/6103.aspx
47

[12] Strengthening Community College Institutional Research and Information,
Public Agenda, 2011
[13] 李紋霞, 符碧真, 全球視野在地化的校務研究：教育科學研究期刊, 第六十
二卷第四期, 2017
[14] Online transaction processing,
https://en.wikipedia.org/wiki/Online_transaction_processinghttps://en.wikipedia.
org/wiki/Online_transaction_processing
[15] Khadija Letrache, Omar El Beggar and Mohammed Ramdani, “OLAP cube
partitioning based on association rules method”, Applied Intelligence , 2019
[16] Data Warehouse, https://en.wikipedia.org/wiki/Data_warehouse
[17] Data Warehousing - Multidimensional OLAP, Tutorialspoint,
https://www.tutorialspoint.com/dwh/dwh_multidimensional_olap.htm
[18] Jiawei Han, Micheline Kamber and Jian Pei, Data mining concepts and
techniques 3rd, Morgan Kaufmann, 2011
[19] Clint Fontanella, A Beginner′s Guide to Customer Behavior Analysis, HubSpot
Blog, https://blog.hubspot.com/service/customer-behavior-analysis
[20] Bayes′ theorem, https://en.wikipedia.org/wiki/Bayes%27_theorem
[21] Jiuyong Li, Thuc D le, Lin Liu and Jixue Liu,“Mining Causal Association
Rules”, Proceedings of the 2013 IEEE 13th International Conference on Data
Mining Workshops, 2013
[22] IBM Knowledge Center, https://www.ibm.com/support/knowledgecenter/zh-tw
[23] Star and SnowFlake Schema in Data Warehousing, Guru99,
https://www.guru99.com/star-snowflake-data-warehousing.html
[24] Data Warehousing - Schemas, Tutorials Point,
https://www.tutorialspoint.com/dwh/dwh_schemas.htm

指導教授

蔡孟峰(Meng-Feng Tsai)

審核日期

2019-7-26

推文