Paper Key : IRJ************796
Author: Bhure Meet Nileshbhai,Dr. Arpit Solanki
Date Published: 04 Apr 2025
Abstract
The real estate market in India is dynamic, with house prices reflecting economic conditions and influencing investment decisions. This research presents a comparative analysis of machine learning (ML) modelsLinear Regression (LR), Naive Bayes (NB), and K-Nearest Neighbors (KNN)to predict house prices in real-time using a dataset of 68,613 test entries and 28,000 training entries from various Indian cities. The study evaluates the impact of features such as size, location, and price on sales predictions, employing regression techniques due to the continuous nature of the target variable. Pre-processing methods enhance prediction accuracy, and performance is assessed using metrics like Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared values. Results indicate that KNN outperforms LR and NB in handling high signal-to-noise ratio data, achieving superior accuracy. This work provides a robust framework for developers and buyers to estimate house prices, addressing biases inherent in traditional appraisal methods.