UCLA

Rental price prediction (price recommendation) is a practical topic in the current online marketplace. In order to support hosts with less experience to set up the competitive rental prices, we utilize the techniques, such as feature engineering and machine learning algorithm, to select useful features and conduct models to predict possible rental prices based on the property information provided by the hosts. In this paper, machine learning algorithms are implemented on the same dataset which contains all the properties in California listed on Airbnb. After feature analysis, we notice the number of bedrooms and property types are the most important features that are highly associated with the rental prices. Among all the methods, XGBoost gives the most satisfying prediction results of rental prices based on RMSE, MAE, and R squared.