Evaluating the Accuracy of Machine Learning Models in Predicting Real Estate Prices: A Comparative Study
(ML )Machine Learning, (ANN) Artificial Neural Networks,( RF) Random Forest, (LR Logistic Regression, (ROC-AUC) Receiver Operating Characteristic Area Under Curve, (SHAP )SHapley Additive exPlanations, (DNN Deep Neural Networks, XGBoost - Extreme Gradient Boosting, (MRT )Mass Rapid Transit, (CFS )Correlation-based Feature Selection,(BFS)Best First Search, (IQR) Interquartile Range
This study explores the application of machine learning (ML) algorithms in predicting real estate prices using housing-related features such as house age, distance to MRT stations, and number of nearby convenience stores. Four models--Logistic Regression, Random Forest, and Neural Networks--were evaluated using a dataset of 414 housing transactions. Performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC were used to compare the models. The goal is to determine the most effective algorithm for house price prediction based on various evaluation metrics and dataset characteristics. The results show that ANN outperforms other models in prediction accuracy, while Random Forest offers an optimal balance between performance and interpretability. SVM demonstrates solid performance on smaller datasets but lacks scalability. Logistic Regression, though efficient, struggles with nonlinear relationships. This research contributes valuable insights for real estate stakeholders aiming to leverage ML for data-driven property valuation.
Atharv Gautam. "Evaluating the Accuracy of Machine Learning Models in Predicting Real Estate Prices: A Comparative Study".INTERNATIONAL JOURNAL OF CREATIVE RESEARCH THOUGHTS ISSN:2320-2882 , Vol.13, Issue 5, pp.q143-q153, URL :https://orwave.org/uploads/ORWave_72.pdf
Volume 13 Issue 5
Pages. q143-q153