Housing Price Prediction

Before we proceed

If you are not able to view the ipynb file through github, try using this nbviewer version

Goal:

To identify the best model that could predict house prices in King County using data provided.

Data:

https://www.kaggle.com/harlfoxem/housesalesprediction

Approach:

• Analyzed the features for potential cleansing
• Split the data into training and test samples
• Applied the following models and identified r2 and MSE values in the training set
  • Linear Regression
  • Lasso
  • Ridge
  • Decision Tree
  • Random Forest
  • XGBoost
• Used hyperopt for hyper parameter tuning
• Predicted using the models and derived r2 and MSE values
• Created a Residual plot for each prediction.
• Identified XGBoost as the clear winner based on all results