house-price-prediction

A House Price Prediction project

This is a project undertaken from the book Hands-On Machine Learning with Scikit-Learn & TensorFlow CONCEPTS, TOOLS, AND TECHNIQUES TO BUILD INTELLIGENT SYSTEMS by Aurélien Géron.

It uses data from the California Housing Prices dataset from the StatLib repository Based on Supervised Learning, it predicts the price of a house in Carlifonia using the number of rooms, median income, ocean procimity etc.

Methods and Technologies used

Python language and its libraries such as Numpy, Pandas, Matplotlib and mainly Scikit-Learn were deployed.

I tested both Random and Stratified Sampling methods for splitting training and testing data but settled on the latter.

For handling text and categorical attributes from the dataset, I used the LabelBinarizer method provided by Scikit-Learn.

A transformation pipeline was fed using a custom transformer and Scikit-Learn's transformers for preparing the data to be fed to the models.

Three prediction methods were tested at first:

• Linear Regression
• Decision Trees
• Random Forest

I settled on the Random Forest method because it had the lowest Root Mean Squared Error.

Grid Search Optimization was used to optimize the model in which various hyperparameters were used.

The code will be improved time to time and OOP will be used. You can also uncomment relevant lines.