Roadmap

Hall is currently in the alpha phase and under development. At the moment, only basic random variable algebra and a handfull of distributions is implemented.

My goal is to make Hall a reliable library that can be used in production environments, as well as a reliable tool to teaching probability theory.

In upcoming releases, hall will add support for:

• Many of the commonly used discrete and continuous probability distributions
• Symbolic algebra of multiple random variables/vectors (without sympy), e.g., the birthday paradox is P(B1 == B2) where B1 and B2 are ~Uniform(1, 365).
• Joint distributions and random vectors
• Conditional random variables, covariance, correlation, etc.
• Complex distributions
• Clean interface for user-defined distributions

With lower probability, hall could also be able to:

• Statistical inference with symbolic parameters (likelihood-based, bayesian, etc.)
• Confidence intervals, with e.g. t- and z- tests.
• Optional numpy support (random matrices, and linear algebra thereof)
• Optional pandas support
• Optional numba or cupy support (for e.g. convolutions) to increase performance

And with an even lower probability, hall might include support for:

• Stochastic processes, maybe even random fields
• Bayesian networks
• Mixture models
• Markov kernels
• Fuzzy logic
• Fuzzy sets
• Possiblity theory