📔 Jupyter-Notebook-Template

Notebook Templates for quick prototyping of Machine Learning solutions.

What's in the project?

STEP 1️⃣: Data Exploration

• EDA using Lux: A Python API for Intelligent Visual Discovery
• Pandas Profiling

STEP 2️⃣: Feature Engineering

• Outlier Treatment
• Missing Values Imputation
• Encoding Categorical Attribute
• Scaling
• Handling Class Imbalance
• Dimensionality Reduction

STEP 3️⃣: Feature Selection

STEP 4️⃣: Modelling

• Classification using Pycaret
• Regression using Pycaret
• Clustering using Pycaret
• Anomaly Detection using Pycaret
• Time Series Modelling
• Recommendation System
• Optimization Problems

STEP 5️⃣: Hyperparameter Optimization

• Cross Validation using Scikit-Learn
• Search Techniques using Scikit-Optimize

STEP 6️⃣: Model Explanability using SHAP

TASKS

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.