Download

Download Instructions

1. Download the Bayesian Modeler and install the setup program (Windows only).
   
2. Email us for a free trial license. Please describe how your application or project can utilize Bayesian Networks.
3. Copy the trial license to your Windows directory (e.g. C:\WINNT, C:\Windows) and rename the file to deriveit.lic
4. Run it from Start-DeriveIt-BayesianModeler Menu.
5. Email us again on how we can work together for your application or project.

Features

1. Exact Inference: 1) Junction tree propagation, 2) Variable elimination, 3) VE1 for causal independence.
   
2. Approximate Inference: 1) MCMC with Gibbs sampling.
   
3. Learning: 1) Learn parameters from complete data - ML or MAP, 2) Learn parameters from incomplete data - EM, 3) Learn structure from complete data, 4) Learn structure from incomplete data - Structure EM.
   
4. Model Scoring: 1) Log likelihood, 2) Conditional log likelihood, 3) Bayesian information criterion, 4) Marginal likelihood using Dirichlet priors, 5) Mutual information.
   
5. Structure Searching: 1) Hill climbing search, 2) K2 search.
   
6. Classification: 1) Naive Bayesian classifier, 2) Tree-augmented naive Bayesian classifier - TAN, 3) Bayesian network classifier.
   
7. EM: 1) Missing data, 2) Hidden variables.
   
8. Causal Independence/Functional Dependency: 1) Standard Noisy-OR/AND/XOR/MAX/MIN/ADD functions, 2) Custom function in C++, 3) Scalable data acquisition, 4) Scalable inference and learning via VE1, 6) Larger and denser networks, 7) Hierarchical Bayesian networks.
   
9. Ease-of-Use: 1) GUI - build network interactively, 2) CSV data files

Screen Shots

Still not sure about what you are getting? Here are some screen shots: