Applications of Machine Learning Models to Predict Property Prices

Abstract

The real estate market has always been filled with uncertainty. There are many players and economical factors that continuously affect property prices. For many home buyers, it is hard to estimate the house prices due to the numerous factors involved. The authors of this paper aim to develop an algorithm to help with the estimations. Thus, the research presented in this paper aims to develop and test machine learning models, then analyze the results of these models as applicable to the housing market. We studied and worked with two datasets, and developed numerous models to apply the best model for the data. We evaluated the results of the models and compared them to each other to draw inferences and conclusions. The results for one dataset were better compared to the results for the other dataset. Similarly, the predicted property prices were closer to the actual prices for one dataset than for the other dataset. We observed multiple linear and non-linear trends within these datasets by closely studying the results. Throughout our study, we have found that machine learning can be used to predict property prices with varying degrees of effectiveness. The results mainly depend on the quality of the data, and how well they reflect the values of the people.