Simple Bayesian Networks with Ruby
summary
A Ruby library called SBN implements simple Bayesian networks, using a Markov chain Monte Carlo approximate inference engine (MCMC) rather than exact inference and learning parameters from sample points. The demo recreates the sprinkler/rain/grass network and reports posterior shifts (grass wet → cloudy rises, rain more likely than sprinkler), claims the Ruby rewrite runs faster than an earlier C++ version, supports XML BIF export, string and numeric variable helpers, and leaves exact inference and continuous vars for future work.
AI-written summary · about
chapters (10)
- 0:00Start
- 9:38robability distribution
- 10:48arginalization
- 14:25conditional
- 22:01esian Networks
- 26:32Cloudy
- 36:21ing probabilities
- 38:46restoring a network
- 39:30ced variable types
- 42:13l-world example
related talks
29mModeling the World Probabilistically using Bayesian Networks in Clojure
41mBayesian Network Modeling using R and Python
30mProbabilistic modelling with Scala
33mExplainable causal inference results
01h 03mAn Overview of Probabilistic Programming
55mHigh-performance functional bayesian inference in Scala
04mINTERVIEW
25mScaling Bayesian Experimentation