An introductory study on the urban functions using CHAID technique

CHAID 技法에 의한 都市機能의 試論的 硏究

To this day, a number of quantitative analytical methods have been employed in clarifying regional characteristics in the discipline of geography. This paper attempted, as a part of application of those quantitative analyses, to make clear the urban functions and consequently the urban characteristics statistically by adopting newly-introduced CHAID, a sort of discriminant analyis technique. The processing of data was sonducted in two phases. To begin with, the urban functions were classified after designating twenty cities - the population of each city counting 250, 000 or more - as predictor variable, and at the same time four major urban functions like administration, marketing, finance and production as response variable. And then, preeminent functions of individual region were discriminated and concurrently classified by treating the remaining traffic, education, medicare, culture and transportation functions as predictor variable, and the following five regions as response variable: Metropolitan Seoul Area. Pusan region, Taegu region, Kwangju region and Chungcheong region. According to the result of this analysis, marketing and administration are emereed as meaningful functions in Seoul and Taegu respectively. As for the finance function only Pusan and Pucheon can be discriminated. Seoul, Pusan and Seongnam reveal their dominancy in production function. To take a look at the result of the latter analysis, the Metropolitan Seoul area shows, among other functions, strong traffic and finance functions. When it comes fo Pusan region, adminstration, education and finance functions are recorded as a leading ones, and Taegu region is preferable in education, medicare and transportation functions. In case of Kwangju region adminstration, production and education functions are discriminated from any other functions. Chungcheong region shows similar aspect with only traffic function replacing the production function of Kwangju region. Based on aforementioned anlysis, it can be said that the CHAID technique, which is capable of processing large amount of categorical data and, by presenting its outcome in the form of dendrogram, facilitates the interpretation work, is an effective, meaningful means to classify and discriminate certain geographical regions and their characteristics.

地理學에서는 地域의 特性을 규명하고자 수많은 計量的 分析手法을 사용하여 왔다. 본 고에서는 일종의 判別分析技法으로 최근에 도입된 CHAID技法을 사용하여 都市와 都市 機能에 관한 통계처리를 시도하였다. 2종류의 자료를 가지고 두 차례 처리를 실시하였는데, 하나는 인구 25만명 이상의 도시 20개를 예측변수로 하고, 行政, 市場, 金融機能 그리고 生 産機能을 반응변수로 하여 도시의 기능을 분류해 내었다. 두번째 처리에서는 앞서 언급한 행정, 시장, 금융, 생산기능 이외에 交通, 敎育, 의료, 文化, 그리고 運送機能의 9가지를 예측 변수로 선정하고, 수도권, 부산권, 대구권, 광주권, 충청권의 5개 권역을 반응변수로 하여 각 권역에서 탁월한 기능을 판별, 분류해 내었다. 이상에서 CHAID기법은 큰 양의 범주형 자료 를 처리할 수 있고, 樹形圖로 결과를 산출하여 해석이 용이하므로 地域을 分類하거나 특성 을 判別하는데 유용한 또 하나외 새로운 분석틀로 여겨진다.