Comparative Study on Artificial Intelligence Methods in Housing Price Prediction

Abstract

The demand for property, including houses, continues to grow rapidly in Indonesia. The housing price prediction is essential in assisting the stakeholders such as buyers, sellers, and investors to make better decision-making. There are many key factors that influencing the housing prices and it is challenging to identify the most relevant factors. This study provides a comparative analysis of various methods in the housing price prediction that consists of one traditional method, Linear Regression (LR), and three artificial intelligence (AI) methods, including Artificial Neural Network (ANN), Classification and Regression Tree
(CART), and Chi-Squared Automatic Interaction Detection (CHAID). The aim is to find the best machine learning method in predicting the housing price in terms of prediction accuracy through the four performance indicators and one combined performance index called the reference index (RI). The main findings of this study is that the AI-based method, the ANN method, has the best accuracy indicated by its highest RI value hence outperforming other methods in predicting the housing
prices.