Comparing Machine Learning Algorithms

• It's important to define which algorithm gives better performance of the machine learning solutions, we need to narrow down on the best algorithms that suit both the data and the business requirements.

• You can see:

  • OOP implementation of the same project
  • Kaggle Version of the same project

Procedures:

• Phase I
  • Data Pre-processing
    • Gathering the Data
    • Data Exploration and Analysis
    • Data Cleaning
      • Outliers Detection
      • Dealing with Null values
      • Adding new features
    • Prepare the data for ML
    • Dealing with the high skewness of the data
• Phase II:
• Initialize the models.
• Comparing different Machine learning models.
  • Compute train/test results.
  • Evaluate our models using cross validation.
• Improving the top models
• Stacking the best models to get a better score

Results :

• Phase 1
  • Nulls

• Skewness

• Phase 2

  • Train/Test

    • R-Squared and Adjusted R Squared

    

    • MAE, MSE, and RMSE

• Cross-Validation

  • R-Squared and Adjusted R Squared

    

  • RMSE

    

Used Data :

House Prices

Algorithms :

• Elastic Net
• Kernel Ridge
• Lasso
• Random Forest
• SVM
• XGBoost
• LGBM
• Gradient Boosting