Tutorials

Step-by-step walkthroughs that build understanding from the ground up. Each tutorial is a runnable Jupyter notebook.

• Important Datastructures
  • Sets
    • Discrete Sets
    • Continuous Sets
  • Variable Assignments
• Univariate Probability Spaces
  • Discrete Domains & Distributions
    • Definition
    • Manipulating Distributions
    • Learning Distributions from Data
  • Continuous Domains & Distributions
    • Real-valued Sets
  • Symbolic Domains & Distributions
• Random Variables
• Variable Assignments
• Working with Intervals
  • Imports
  • ContinuousSet
    • Creating intervals
    • Parsing from strings
    • Basic properties
    • Predefined constants
  • Set Operations
    • Complement
  • Sampling
  • IntSet — Integer Intervals
  • UnionSet — Disjoint Unions
    • Simplifying a UnionSet
  • Chopping Intervals
  • Transforming Boundaries
  • Using Intervals in JPT Queries
    • Evidence as intervals
    • Posterior domains are intervals
    • Building multi-range evidence with UnionSet
• Learning of Joint Probability Distributions
  • Preparing the Data
  • Inferring Variables from a DataFrame
  • Creating and Fitting a JPT
  • Effect of Hyperparameters
  • Generative vs. Discriminative Learning
  • Saving and Loading a Model
• Reasoning about Joint Probability Distributions
  • Marginal Queries
  • Conditional Queries
  • Posterior Distributions
  • Expectations
  • Most Probable Explanation (MPE)
  • k Most Probable Explanations (k-MPE)
  • Using bind to Construct Queries
• Visualising JPTs
  • Setup
  • Tree Structure
    • Restricting which variables appear in leaves
  • Plotting Individual Distributions
    • Numeric distribution — CDF
    • Symbolic distribution — bar chart
  • Posterior Distributions
    • Numeric posterior — CDF
    • Numeric posterior — PDF (manual plot)
    • Symbolic posterior — bar chart (Plotly)
  • Plotly Engine — Numeric CDF
  • Customising Colours
  • Saving Figures
    • Static images (Matplotlib)
    • Interactive HTML (Plotly)
    • Static images (Plotly + kaleido)