Pembangunan Sistem Geodemografi di Pulau Pinang:Proses Pemilihan Variabel dengan MenggunakanAnalisis Komponen Utama (PCA)

Development of Geodemographic System in Penang:Variable Process Selection by Using thePrincipal Component Analysis (PCA)

Abstract

In general, geodemographics can be defined as the study of people ang its realtion to where they live. One of the thrusts in this field is area classification which involves major components such as digital data, data mining and geographic information system. The main source of digital data for the development of geodemographics system is census data which involves comprehensive data collection of demographic information in a particular area. For example, the data base of population and housing census developed by the Department of Statistics in 2000, has more than 190 variables that can be used as inputs to develop this classification system. However, this large amount of variables can not be used as input to develop a classification system due to various problems. Thus, variable selection process has been carried out in advance to ensure the variables used in the cluster formation do not contain repeated information. One of the variable selection methods that can be used is the Principal Component Analysis (PCA), which divides the census data into small groups. Apartfrom dividing variables into separate components, the PCA analysis can also be usedto select individual variables based on Eigen value produced.