諮詢系統自我評量能力的建立與探討

Establish Self-Assessment Consultant System by Data Mining Approach

曾伊禪(Yi-Chan Tseng)

指導教授 ： 蔣定安

淡江大學/工學院/資訊工程學系碩士在職專班/碩士(2007年)

https://doi.org/10.6846/TKU.2007.00487

摘要

評量表是一種可以快速了解，我們輔導對象目前狀況的一個工具，而目前評量大部分均以固定型式問卷來表示，這樣評量表不但較為空洞泛味，而且也常常流於形式，而大大的降低評量表的功效為了要讓問卷可以更有效發揮功能，縮短問卷問答題數或者加強重要題目的比重是必要的，為了達到這個目的我們期望，可以透過資料採礦的技術與問卷結果的搜集，進而分析計算出一份問卷題型最佳組合，為了要計算出最佳組合，我們透過IM8(Intelligent Miner for Data 8.2)將資料以決策樹(Decision Trees)及關聯式分析
(Association riles)分別做出相關分析，對於問卷我們先將問卷題目以分類向做一群比，更把整份問卷的回答過程模擬為樹的尋訪根據資料採礦的分析結果作為樹尋訪時，選擇節點的依據，如此一來便能以最佳路徑的演算法得一個最佳的組合。
關鍵字：決策樹、關聯式法則、決策分析、自我評量表

關鍵字

決策樹 ； 關聯式法則 ； 決策分析 ； 自我評量表

並列摘要

Assessment sheet is one kind of tools to quickly understand the present situations of the counseling object. Most assessment sheets are presented by the fixed questionnaire nowadays; however, it is a tedious and modal way. This would lower the efficiency of the assessment. In order to let the questionnaire have the most efficiency, it is necessary to decrease the number of questions or increase the proportion of the important questions. To achieve the target, we expect to use the technique of Data Mining and the collection of the poll results; then to analyze and obtain the optimized combination of assessment sheet. We finished the related analysis by using decision trees and association riles methods through IM8 (Intelligent Miner for Data 8.2). We first group the assessment sheets by the subjects and then simulated the filling procedure to be the searching base of tree. The analysis result of Data Mining is used to be the reference of selection nodes when searching the trees. Consequently, the optimized combination is obtained with the algorism of optimized path.

並列關鍵字

Decision Trees ； Association rules ； Decision analysis ； Self-Assessment Table

參考文獻

李延震，”資料探勘-學習與讀書策略”，淡江大學資訊工程學系研究所碩士論文，2006。

Berry, L.L. & Linoff, G. S. (2002). Mastering Data Mining, The Art & Science of Customer Relationship Management. NY: John Wiley & Sons.

Todman, C. (2001). Designing a Data Warehouse: Supporting Customer Relationship Management. Upper Saddle River, NJ: Prentice Hall.

S. Chakrabarti. Mining The Web: Discovering Knowledge From Hypertext Data.

J. T. Wijnen, H. F. A. Vasen, P. M. Khan, A. H. Zwinderman, H. Van Der Klift, A. Mulder, C. Tops, P. Moller, and R. Fodde. Clinical ‾ndings with implications for genetic testing in families with clustering of colorectal cancer. N. Engl. J. Med., 339:511{518, 1998.