Machine learning-based forecasting of residential property prices in Mumbai

This analysis uses machine learning models to make price predictions of residential properties in Mumbai based on data available in the Kaggle platform in the dataset, Mumbai House Prices. The data record has over 76,000 records that feature 9 attributes of properties and both categorical and numerical data points. The analysis employed several regression models that comprised Linear Regression along with Support Vector Regression (SVR), K-Nearest Neighbors (KNN), Random Forest, and XGBoost. To achieve the best performance of the model, it was necessary to apply data cleaning, one-hot encoding, and feature scaling processes. The performance of the results was assessed through Mean Absolute Error (MAE) along with Mean Squared Error (MSE) and Root Mean Squared Error (RMSE), and R² Score. Random Forest model demonstrated the best predictive performance with an R² score of 0.9058 during testing and proved its superiority and ability to generalize well. KNN parameter tuning caused it to exhibit an impressive performance, which shows that proper parameter tuning can cause the simple algorithms to rival the complex ones in terms of prediction accuracy. The choice of models and optimization of their hyperparameters, based on the results, play a crucial role in predicting real estate prices.