Real Estate Price Prediction Using Machine Learning: A Comprehensive Web-Based Prediction Platform

Abstract

Accurate real estate price prediction is critical for buyers, sellers, investors, and policymakers in today's dynamic property markets. This paper presents a comprehensive web-based real estate price prediction system using machine learning algorithms, implemented as a Flask Python application accessible through standard web browsers. The system integrates multiple regression algorithms—Linear Regression, Decision Tree, Random Forest, and XGBoost—with a complete data preprocessing pipeline, interactive model training interface, and feature importance visualization. The architecture comprises five integrated modules: (1) User Interface Module providing intuitive web-based data upload, model training, and prediction interfaces; (2) Data Processing Module implementing automated cleaning, missing value handling, categorical encoding, and feature scaling; (3) Machine Learning Module supporting four regression algorithms with configurable hyperparameters; (4) Prediction Module generating price estimates with feature importance explanations; (5) Visualization Module rendering performance metrics, actual vs. predicted comparisons, and feature importance charts. The system was evaluated on real estate datasets using 80-20 train-test splits and 5-fold cross-validation. Experimental results demonstrate XGBoost achieving the best performance with R² = 0.89, Mean Absolute Error of ₹10,850, and RMSE of ₹13,850, outperforming Linear Regression (R² = 0.86), Decision Tree (R² = 0.82), and Random Forest (R² = 0.88). Comparative analysis shows the proposed system achieves 6% higher R² than commercial platforms such as Zillow Zestimate (R² ≈ 0.84) while providing user customization, transparency, and open-source accessibility that proprietary systems lack. User acceptance testing with 20 participants yielded an overall satisfaction rating of 4.5/5.0. This work provides an accessible, transparent, and customizable platform for real estate valuation that democratizes advanced machine learning analytics for non-technical users