Mining Patterns in Housing Characteristics of Migrants

摘要

Clustering deals with the investigation of information in a collection of unlabeled dataset. It is also one of the vital technologies in analyzing information for a country's population and housing data. As aspect of demography, human migration is one of the most issues for country development and, if managed effectively, it brings benefit to the host and destination areas. Migrants' demographic factors are crucial parts to be considered for country's development. This study aims to investigate housing characteristics of migrant's households. This paper discovers patterns of different migrants' housing features by applying clustering techniques and evaluates these patterns using validity measures and cluster visualization. This study applies five clustering methods for analyzing and obtaining an appropriate method for dataset. In addition, the study implements a web application for visualization and evaluation of obtained patterns from clustering methods so it allows users to use and compare the clustering results with ease.