Notebook Templates for quick prototyping of Machine Learning solutions.
- Outlier Treatment
- Missing Values Imputation
- Encoding Categorical Attribute
- Scaling
- Handling Class Imbalance
- Dimensionality Reduction
- Classification using Pycaret
- Regression using Pycaret
- Clustering using Pycaret
- Anomaly Detection using Pycaret
- Time Series Modelling
- Recommendation System
- Optimization Problems
- Cross Validation using Scikit-Learn
- Search Techniques using Scikit-Optimize
STEP 6️⃣: Model Explanability using SHAP
TO-DO
- EDA using Autoviz
- Human Explanation using EDA
- TimeSeries Modelling
- Recommendation System
- Optimization Problem
- Clusterig - Extend the models list, currently pycaret clustering module is used which is limited.
- Explanaible AI
- Neural Networks
- Demensionality Reduction
- Optimise Thresholds for classification - Refer here for details.
- Model Monitoring - Refer here for details.
- Review and add if any Feature engineering step missed using github repo here.
- Add Multiple Time Series Forecasting method from pycaret, refer here.
- Include Halving Grid Search(20X times faster)
- Include Market Risk code from code reference folder.
- https://github.com/vaexio/vaex/
- ChefBoost package
- https://towardsdatascience.com/a-python-library-to-remove-collinearity-5a4eb05d3d73
- Explore AutoML frameworks here - https://www.kaggle.com/rohanrao/automl-tutorial-tps-september-2021
COMPLETED
- EDA Using LUX
- Feature Engineering using Scikit-Learn and ImbLearn
- Classification Template using Pycaret
- Regression Template using Pycaret
- Clustering using Pycaret
- Anomaly Detection using Pycaret
- Hyperparameter Optmization using Scikit-Learn & Scikit-Optimize
- Feature Selection - Refer here for details.
Resources are mentioned in more details in document here.