Comparing Machine Learning Techniques for House Price Prediction

Abstract

One sector that already is feeling the impact of Artificial Intelligence (AI) is the real estate industry. AI is being used in the real estate sector to improve various aspects of the industry such as property search, pricing, marketing, and risk management. There are various techniques used for predicting house prices, including linear regression, decision tree and random forest. These algorithms can take into account factors such as location, square footage, number of bedrooms and bathrooms, and other relevant characteristics of the property. Predicting house prices with AI has many advantages over traditional methods, such as the ability a) to handle large amounts of data and b) to identify patterns and trends that might be overlooked by humans. It can also help us understand why housing prices are changing, and what factors are driving these changes. This information can be invaluable for real estate agents, investors, and homeowners who need to make informed decisions regarding the market. This paper provides a comparison of the performance of various Machine Learning algorithms in their attempt to predict the price of houses. The regression methods that are compared include Support Vector Machine, Kernel Ridge, Gradient Boosting, Lasso, Random Forest, XGB, LGBM, Average and Voting Regressor. The comparison shows that the best algorithm was Voting Regressor for R Squared metric and for RMSLE metric was the Average model. In conclusion, AI has the potential to bring new levels of accuracy and insight into the real estate industry.