Housing-Price-Prediction

Kaggle Housing Prices project — ML pipeline with preprocessing, feature engineering, and regression models

Housing Price Prediction 🏡

Overview

This project is based on the Kaggle "House Prices: Advanced Regression Techniques" dataset.
The goal is to predict house sale prices using machine learning models and feature engineering.

Dataset

• Source: Kaggle House Prices Challenge
• Includes features such as lot size, year built, number of rooms, and neighborhood.

Approach

1. Data Preprocessing

   • Missing value imputation
   • Feature scaling and encoding
   • Handling multicollinearity

2. Feature Engineering

   • Created new features (e.g., age of house, total square footage)
   • One-Hot Encoder and Ordinal Encoding together inside a ColumnTransformer / Pipeline is done because different types of categorical features transformation.
   • Selected important predictors

3. Modeling

   • Tried Linear Regression, Random Forest, and XGBoost
   • Evaluated using RMSE

4. Results

   • Best model: The RandomForestRegressor achieved a test RMSE of 0.14 in the log(Sale Price) scale, which translates to an average prediction error of approximately 15% in the original sale price units. For example, for a property valued at £200,000, the model’s prediction would typically be within ±£30,000 of the actual price. The R² score of 0.87 indicates that the model explains 87% of the variability in housing prices, demonstrating strong predictive performance.

   • Insights: Feature importance shows neighborhood and square footage as key drivers.

How to Run

1. Clone the repo:

   git clone <https://github.com/Adibab/Housing-Price-Prediction.git>