Evaluation of Price Prediction of Houses in a Real Estate via Machine Learning

Abstract

Traditional (manual) methods of determining real estate house prices are in some cases prone to mistakes which may be due to distractions, lack of attentiveness or vulnerability to real estate agent fraud. This work focuses on evaluation house price prediction in real estate using more recent methods. House pricing using such methods as House Pricing Index and Random Forest Machine Learning Technique has been discussed, a new approach is proposed as a model utilizing the Extra Tree regression because it introduces an additional level of randomness in the tree-building process. Kaggle Boston housing dataset with 506 entries and 14 features was employed to train and test the developed model whose efficiency was then determined via mean absolute error and mean squared error. Additionally, a comparison was made between a random forest regression model and the proposed prediction model which revealed that the new prediction model yielded better performance than the random forest regression.