The housing market is one of the earliest and most influential industries with interests among general populations. In recent years and with the advent of computer approaches, many studies used the latest machine learning models to analyze the housing market and identify its most important influential variables in order to suggest a proper price or to predict price fluctuations. This article follows the general phases of the CRISP-DM process model for data mining to elaborate on the problem statements, data collection and preparation, modeling, and evaluation. It suggests proper ways to design steady and accurate models in relation to previous methods and approaches for predicting housing prices. Based on this investigation, previous methods suffer from reaching steady results on multiple datasets, which can be largely attributed to the existence of bias in training, as it is essential to predict prices by considering external economic variables.
