Application of Feature Engineering Techniques and Machine Learning Algorithms for Property Price Prediction

Denny Jean Cross Sihombing
DOI: https://doi.org/10.62527/jitsi.5.2.241
2024-06-30
Abstract:This research applies feature engineering techniques and machine learning algorithms to predict property prices using a dataset from Kaggle. Three models were implemented: Linear Regression, Decision Tree, and Random Forest. The Random Forest model demonstrated the best performance with an average Mean Absolute Error (MAE) of 16472.76, Mean Squared Error (MSE) of 457407807.78, and R-squared (R2) of 0.83. Key features influencing property prices were identified through feature importance analysis, providing valuable insights for enhancing property appraisals and investment decisions.
What problem does this paper attempt to address?