以聯招考生選填志願挖掘系所關聯性之研究

A Study of the Competition relationship among Departments by Mining Association Rules between Departments in Eaxminee Future Intenton Databases of the College Entrace Examination

楊哲智(Che-Chih Yang)

指導教授 ： 蔡介元

元智大學/工程學院/工業工程與管理學系/碩士(2004年)

摘要

近年來，隨著出生人口下降及大學院校招生名額大幅增加，已使國內高等教育市場飽和，各大專院校面臨的招生競爭也愈來愈激烈。在飽和招生市場中，市場競爭將以「顧客導向」或「競爭者導向」的方式進行，因此，各校系必需由考生的角度來瞭解其主要競爭對手及市場定位。然而，各校系由於缺乏適當的方法了解其產品定位及主要競爭對手，以致無法針對目標市場全力發展。因此，本研究目的為發展一個考生選填志願關聯性分析方法與系統，由考生選填志願後的資料發覺系所的關聯性，以避免研究方法受到人心理因素的影響，並從考生的角度來瞭解各校系之市場供需情況、產品定位及主要競爭對手。由於考生選填志願的資料型態與商品交易型態非常類似，利用關聯法則由龐大的考生選填志願資料中找出各校系的關聯性。本研究改善IHP演算法為MSbinary演算法，順利完成考生選填志願關聯性分析系統。分析92年大學聯招87059位錄取考生選填志願資料，得知全國校系產品定位，以逐步篩選法得知各系所之主要競爭對手。發覺音樂類及體育類校系招生名額過剩，全國共47個校系沒有主要競爭對手，各系支持度受選用「案別」及學校地理位置的影響。以台灣大學、聯合大學、宜蘭大學為例，示範學校關聯性分析方法，顯示台灣大學之主要競爭對手為清華大學，聯合大學之主要競爭對手為宜蘭大學。以聯合大學經營管理學系為例，示範系所關聯性分析方法，顯示聯合大學財務金融學系為其主要競爭對手。

關鍵字

資料挖掘 ； 關聯法則 ； 大學聯招 ； 目標市場

並列摘要

As the dropping of birth rate and the continuous increasing of college enrolled students, the higher education market had become saturated in Taiwan. Every college is confronting the most enormous pressure on recruiting the students. In the saturated market, Market Orientation has become Customer-Oriented or Competitor-Oriented. So that, every college should not only understand its product position, but also its major competitors from the view of customers. However, because of short of appropriate method to understand, school authorities usually can’t do their best on their target market. Examinee Future Intention Databases of National College Entrance Examinations (EFIDof NCEE) are similar to transaction databases. So that, association rule mining could be used in EFIDofNCEE as in transaction databases. The aim of this paper is to develop a system that can help them to understand their supply and demand, product positions and major competitors from the view of customers by mining association rules between departments in EFIDofNCEE beyond the psychological factors. This paper carried out the aim by MSbinary algorithm which is modified from IHP algorithm proposed by Holt and Chung (2002). In this paper, the major competitors of each department had progressively sieved out by MSbinary algorithm from 87,059 freshmen transactions in 2003 EFIDofNCEE. We find that the allowance of music and physical education is too much. There are 47 departments had no major competitors. The support of some departments is influenced by their enrolled mode and school geographical location. We analyzed National Taiwan University (NTU), National United University (NUU), National Ilan University (NIU) and the department of Business Management, NUU (BMNUU) as examples. We found the major competitor of NUU is National Tsing-Hua University, NUU is NIU each other and BMNUU is the department of Finance, NUU.

並列關鍵字

Data Mining ； Association Rules ； College Entrance Exam ； Target Matket

參考文獻

2.David S. Moore (2002), Statistics: Concepts and Controversies, 5th edition, 鄭惟厚譯，「統計學的世界」，一版，天下遠見，台北，pp. 356-369。

15.劉泓郁 (2003)，「建構一個以模糊關聯法則為基礎之產品開發系統」，碩士論文，元智大學工業工程與管理研究所，民國92年7月。

17.Agrawal, R., T. Imielinsld and A. Swami (1993), “Mining Association Rules Between Sets of Items in Large Databases,” Proc. of the ACM SIGMOD International Conf. on Management of Data, pp. 207-216.

19.Berzal, F., J. C. Cubero, N. Marín and J. M. Serrano (2001),”TBAR: An efficient method for association rule mining in relational databases,” Data & Knowledge Engineering vol. 37, pp. 47-64.

20.Bingham, F. G.. (1993), “Marketing the institution of higher learning: a research analysis enrollment model,” Journal of Marketing for Higher Education, vol. 4, pp. 59-72.