Cookbook

Personal reference scripts for commonly used code

• Machine Learning: A folder containing scripts for commonly used machine learning code
  • Model Training: Training, tuning, and evaluating models
    • Cross Validation
    • Hyperparameter Tuning
    • Optimal Class Probability Cutoffs
    • Prediction Intervals
    • Blending Model Predictions
    • Feature Selection
  • Evaluation Plots: Diagnostic plots for evaluating models
    • Residuals
    • Learning Curve
    • Validation Curve
    • Ensemble Model's Feature Importance
    • Decision Tree Visualization
    • Feature Importance with SHAP (SHapely Additive exPlanations)
  • Miscellaneous: Relevant code that doesn't belong in the other sections
    • Assumption Testing
    • Initial Model Performance Testing
  • Spark ML: Building spark.ml models
    • Data Prepartion
    • Cross Validation
    • Hyperparameter Tuning
    • Model Diagnostics
    • Model Training
• Exploratory Data Analysis & Pre-Processing: Preparing data for machine learning tasks
  • EDA Report
  • Missing Values
  • Outliers
  • Pre-processing
• Natural Language Processing: NLTK natural language processing tasks
  • Tokenizing
  • Removing Stop Words
  • Lemmatizing (note: similar to stemming)
  • TF-IDF
  • Topic Modeling
• Plotting: Code snippets for common plots
  • Matplotlib & Seaborn
  • Plotly
• Deep Learning (CNN): Convolutional neural networks in Keras, CNTK, PyTorch, and TensorFlow
• DevOps: A folder containing scripts for operationalizing machine learning models
  • Flask: Operationalizing a trained machine learning model as a RESTful API
    • Web App
    • Sending Requests