結合群集分析與關聯規則於違章建築管理之研究

Applying Clustering and Association rule Techniques on the Research of Squatter Management

李永山(Yung-Sun Lee)；李惠玲(Hui-Ling Lee)

違章建築 ； 違章建築管理 ； 群集演算法 ； 關聯規則演算法 ； squatter ； squatter management ； clustering algorithm ； association rule algorithm

資訊管理展望

15卷1&2期（2013 / 12 / 01）

1 - 17

繁體中文

本研究從2008-2009年台北市違章建築資料中，彙整各行政區人口密度、房價、受教育程度、工商金融狀況、治安狀況、及自有住宅比率等特性，運用群集、與關聯規則演算法，分析台北市違章建築之關聯性，提供相關單位制定管理決策之參考。研究發現如下：屋頂及陽台為最熱門的違建位置；金屬為最常用的違建材料；違建發生相對較高之三個群集地區，分別為：人口密度較高且房價較高、受大學以上教育人數比率較高且竊盜及犯罪發生率較低、及每萬人享有金融機構多且擁有自有住宅比率較高之行政區；違建發生之重要影響因素為該地區受教育程度、人口密度、工商金融狀況、及房價。

The samples of this research were based on the database of Taipei squatter between 2008 and 2009. We explored six characteristics of the administrative districts, including population density, house price, educational level, commercial states, public security, and private house rate. After data pre-processing, we apply data mining techniques-K-Means clustering algorithm and Apriori association rule algorithm to analysis the relationship between squatter. In this study we found that: the most popular squatter location is on roofs and balconies; the most commonly used material of squatter is metal; the top three group districts with high rate of squatter are the district of higher population density and higher house price, the district of higher education rate, and lower theft and crime rate, and the district that have more financial institutes, and the higher rate of private house. The critical factors of squatter are the education level, population density, the numbers of financial institutes, and the house price.