Hospital Insurance Charges Prediction

Purpose of The Build

Nothing special; curious-based project

Disclaimer: This project is just for builder to experimenting on Streamlit. Don't take it too seriously if there are some mistakes made!

Dataset

Hospital Insurance Charges extracted from Kaggle.

Feature Engineering

1. Delete all duplicates
2. Delete all missing values
3. Analyze outlier and delete as necessary
4. One-hot encoding for all categorical/object type features
5. Standard scaling for all numerical/int & float type features
6. Using TransformedTargetRegressor for easy charges interpretation
   

Analytical Approach

1. Extract Datasets
2. Do feature engineering
3. Search best model by using model benchmarking
4. Search best model parameter by using hyperparameter tuning
5. Build best model and save using pickle
6. Deploy to Github and VOILA!
   

Deployments

You can check my website prediction deployment here

Conclusion & Recommendation

Yooo, it works!!! I don't believe it!



©2024 avrotech