keneandita/hpp repository overview

⁠Project Summary

This repository contains a containerized version of Housing Price Predictor, a machine learning application designed to estimate housing prices based on multiple property features. The project integrates a full data preprocessing and feature engineering pipeline to clean, transform, and prepare data for predictive modeling. Multiple regression models are included, enabling both accurate predictions and interpretability.

The application is suitable for educational purposes, experimental workflows, or lightweight real-world deployment. Dockerization ensures consistent execution across environments without manual dependency setup.

⁠Architecture Overview

Housing Price Predictor is structured around a modular machine learning pipeline:

• Data Preprocessing & Feature Engineering: Handles missing values, encoding categorical variables, scaling numerical features, and constructing derived features to enhance model performance.

• Regression Models: Supports multiple models for prediction, including:

  • Linear Regression
  • Random Forest Regressor Additional models such as Gradient Boosting or XGBoost can be integrated easily.

• Model Evaluation: Uses metrics like R², MAE, and RMSE to assess model accuracy and performance.

• Model Exporting: Trained models are serialized as .pkl files for reuse or deployment.

• Container Environment: Docker encapsulates Python, required libraries, and serialized models for consistent deployment.

⁠Key Features

Data Preprocessing & Feature Engineering

Cleans and transforms raw housing data to maximize model performance.

Multiple Regression Models

Supports Linear Regression and Random Forest Regressor, with easy extension to other regressors.

Model Evaluation

Assesses model performance with R², MAE, and RMSE metrics.

Model Exporting

Trained models are saved as .pkl files for reuse or deployment in other applications.

Containerized Deployment

Docker ensures the application runs consistently without local dependency setup.

⁠Model Performance (Sample)

⁠How to Run

First, pull the Docker image:

docker pull keneandita/hpp

Run the container while exposing the application port (if a web interface or API is included):

docker run -p 5000:5000 keneandita/hpp

After starting the container, access the application (for web or API usage) at:

http://localhost:5000

The containerized setup allows users to predict housing prices, test models, or integrate the trained .pkl models into other workflows without local setup.