This page provides information about what to expect in future releases. The main projects under development are listed below.  In addition, work is ongoing in the areas of optimization, parallelization and 64 bit support.

If you have any feedback on proposed features, or want to register your support for features please let us know.

data sampling	A library for sampling data from Bayesian networks (BayesServer.Data.Sampling), supporting both discrete and continuous variables and Dynamic Bayesian networks.
parameter learning	A library for learning parameters from data (BayesServer.Learning.Parameters), supporting both discrete and continuous variables and Dynamic Bayesian networks.
user interface	User interface for building Bayesian networks, that integrates seamlessly with databases for prediction and learning.
relevance tree algorithm	New inference algorithm which efficiently computes large numbers of marginal distributions.  This is particularly useful for learning Dynamic Bayesian networks.

The screen shot below gives an early impression of the Bayes Server User Interface, which seamlessly integrates your models with data.

The following screen shots show data generated from a number of different Dynamic Bayesian networks using the BayesServer.Data.Sampling library which is currently under development.

The following data was generated by sampling 500 random length sequences from a Kalman filter model.

The following data was generated by sampling 500 random length sequences from a left-right Hidden Markov model.