A Comparative Study of Urban House Price Prediction using Machine Learning Algorithms

Abstract

Accurate housing price forecasts are essential for several reasons. First, it allows individuals to make informed decisions about buying or selling real estate and to determine appropriate prices. Secondly, it helps real estate agents and investors make better investment decisions and negotiate contracts more effectively. In addition, housing prices are often an indication of the general state of the economy. A price decrease may indicate an economic recession, while an increase in prices may signal economic growth. In this study, we proposed to address this subject by predicting house prices using machine learning by choosing three types of machine learning: Linear Regression (LN), Random Forest (RF) and GradientBoosting (GB). We tested our models on the Melbourne real estate dataset, which includes 34,857 property sales and 21 features.