A Data Mining Study on House Price in Central Regions of Taiwan Using Education Categorical Data, Environmental Indicators, and House Features Data

Abstract

This study takes the city of Taichung, Taiwan, as the research area, combines the survey results about the demand for residential houses for the next year, and uses relevant parameters and data of real price registration as the prediction results. In this study, eight types of school district features (such as teachers and students of secondary and elementary schools) and five types of air pollution features are selected and processed with a data mining method to discover the total transactions of real estate properties in various districts of Taichung. The results of K-means clustering and decision tree classification reveal that the four districts of the old Taichung City, namely, Beitun District, North District, Xitun District, and Nantun District, have houses meeting the conditions of egg yolk districts; houses in the old Taichung County have attributes of egg white districts. The results of decision tree classification show that the total price is the most important attribute influencing egg yolk and egg white districts.