본문 바로가기
  • Home

A Study on Apartment Price Models Using Regression Model and Neural Network Model Taehun Kim and Hankuk Hong

Kim Taehun 1 Hong Han Kuk 2

1경성대학교
2동의대학교

Accredited

ABSTRACT

It is in the present situation that there are many studies that estimate and presume apartment price by using various characteristics of apartment. Most of these studies depend on Regression model excessively, but Regression model has more merit than demerit. Thus this study does not deny Regression model but attempts to introduce new model. That is to say, this study is performed by means of the necessity to overcome the problems of Regression model and to introduce complementary model. The main purpose of this study is to presume apartment price by using neural network model of various characteristics of apartment and to compare it with regression model. And examining the complementary aspect of regression model and neural network model is the secondary purpose of this study. In the characteristics of apartment, data easily available in the surroundings are put first in importance. We collected apartment sale price of Songpa-gu and Dobong-gu in Seoul and the characteristics of 12 apartments on the basis of 6/2004. We unified apartment sale prices (that is, sale lowest limit price, general trade price, sale upper limit price) into one sale price by using new measurement method. It is significant in the study field of apartment price to estimate apartment price precisely and effectively by introducing neural network model and to compare it with the existing regression model. It is judged that neural network model can be applied to the past studies and new studies concerning housing.

Citation status

* References for papers published after 2022 are currently being built.