Regression with Playground Series - S4E12
Kaggle Playground December 2024 - Regression with an Insurance Dataset
Goal: The objectives of this challenge is to predict insurance premiums based on various factors.
Evaluation: Submissions are evaluated using the Root Mean Squared Logarithmic Error (RMSLE).
Learnings:
- Log-transform the output to train models
- Feature Engineer with FeatureTools
- Apply Statistic Test in EDA step
- Feature Selection
Table of Contents
1. Preparation
2. Exploratory Data Analysis (EDA)
3. Statistical Analysis
4. Define Metrics
5. Preprocessing (Target Encoder and Standard Scaler)
6. Feature Selection with base models
7. Machine Learning (XGBoost, LightGBM, CatBoost)
8. Hyperparameter Tuning
9. Ensemble Methods (Stacking and Blending)
10. Model Evaluation
11. Predict & Submit
> Achievement: My RMSLE is 1.06
* Check out more details in my [Medium blog](https://medium.com/@nguyenhuynh01.nh/list/kaggle-competition-data-science-b5ae64e91157).
* Check out more details in my Medium blog.
