Predicting US Airbnb Listing Prices by Machine Learning Models

Abstract

This paper addresses the prediction of Airbnb property prices using 2023 open data through the application of machine learning methodologies. In the context of the flourishing sharing economy, accurate price prediction within the short-term rental market holds great significance for hosts and users alike. Drawing on the 2023 Airbnb open dataset, the study employs three distinct models – Linear Regression, Random Forest, and XGBoost. Rigorous training, testing, and evaluation of these models reveal insights into their predictive capabilities. The focus centers on assessing model fit using essential evaluation metrics including R-squared, Mean Squared Error, and Root Mean Squared Error. Results demonstrate that the XGBoost model outperforms both Linear Regression and Random Forest. After parameter tuning, the best parameter for XGBoost regressor exhibits the lowest prediction error and highest R-squared value, showcasing its ability to capture intricate patterns within the data. This outcome underscores the potency of advanced ensemble learning techniques for precise property price predictions. The study's implications are substantial, offering hosts and potential guests improved decision-making insights.