Anglican supports Jupyter-notebook-like browser-based interactive computing that combines REPL (read evaluate print loop) style interactive programming intermingled with Markdown+MathJax documentation.

The worksheet links below are canned, view only workbooks. In order to create your own interactive worksheets or interact with these examples you must follow the browser-based REPL usage instructions.

Practical

• User form filling behaviour

Financial

• Internal Rate of Return 1
• Internal Rate of Return 2

Academic

Pedagogical Models

• Linear Regression
• Classification
• Conjugate Gaussians
• Kalman Smoother
• Gaussian Mixture Model
• Beta-Binomial Model
• Bayes Net

Optimization (requires BOPP and Deodorant)

• Inferring the parameters of a chaotic system

Bayesian Nonparametrics

• Dirichlet Process Mixture Model
• Hierarchical Dirichlet Process
• Probabilistic Deterministic Infinite Automata

Other Examples

• Decision Diagram
• Schelling Coordination Game
• Poisson Execution Traces
• Indian GPA
• Nested Number Guessing
• Marsaglia
• Aircraft Detection

External

• Predicting NYC real estate prices
• Automated tuning of the JVM with Bayesian Optimization

Older Examples

These are the original example models used in the 2014 AISTATS paper introducing Anglican.

• Gaussian with Unknown Mean
• Hidden Markov Model
• A Simple Branching Example
• Marsaglia Normal

Contribute

We would love for you to share the models you develop using Anglican with the community. To share an interesting model follow these steps.

We welcome contributions of new example models and use cases. Please contribute back new and interesting models.