Going the distance: Application of Machine Learning Algorithms in predicting Housing Price Dynamics

In this paper, we have used various algorithms such as Gradient Boost Regression, Catboost Regression, Decision Tree Regression and K-neighbors Regression by categorizing predictions based on threshold. To address the problems of data quality and other complexities, we performed validation on all the algorithms mentioned above. Predicting house prices often involves many barriers such as taking all the important factors into account, being aware of the demand of a particular area, government policies, spatial locations and houses with different sizes. We overcame these problems with the ML models we built. Among all the algorithms, Catboost Regression gave great results with mean absolute percentage error of 0.018 and r-squared value of 0.904. This paper provides insights in more detail about the strong positive correlation between area and price.