A hybrid Regression method for Predicting Housing Prices

Abstract

Accurate house price prediction is crucial for accommodating the diverse needs of stakeholders in the home-buying process. House prices can be affected by various factors, such as location, construction date, exterior, etc. This work proposes a hybrid regression method that leverages the strengths of different regression techniques to improve prediction accuracy. Specifically, this work looks at conventional linear regression and other machine learning techniques such as support vector regression (SVR), and XGBoost regression. Then we compare these models with our proposed hybrid regression model that leverages ridge regression and lasso regression to capture hidden relationships between house properties and sale prices to reveal the different predictive power of these models. In addition, this work also highlights feature engineering to address potential issues in the data and improve prediction performance. The dataset used in this study is obtained from the Kaggle Competition “House Prices: Advanced Regression Techniques.” Different model results are submitted to Kaggle, and the scores are illustrated in the paper.

Keywords: House price prediction, Regression techniques, Feature engineering, Machine Learning

DOI: 10.54941/ahfe1005725