Machine learning on USA house price prediction

Abstract

Nowadays, an increasing number of students are opting to study abroad in order to acquire more advanced knowledge and pursue a superior educational environment. In many foreign countries, the option to apply for school dormitories is only available during the first year of university or graduate school. At other times, international students have to search for rented apartments or apply to stay with local host families. However, when studying abroad for an extended period, purchasing a property can potentially result in significant savings compared to renting. Therefore, this study focuses on comparing three types of machine learning techniques: multiple linear regression, Random Forest, and XGboost in predicting house prices in the United States. This research could provide reference for families studying abroad or property investors. Based on the preliminary findings of this study so far, it can be concluded that the XG-boost model demonstrates the highest accuracy and stability among these three methods.