Software

With its version 3.1, BayesiaLab offers an impressive set of functionalities for modeling and data mining with Bayesian networks:

• Ergonomic graphical interface allowing creating your models by simple clicks, with equation editor for a compact description of your probability distributions
• Complete set of Bayesian networks learning algorithms: Supervised learning algorithms for profiling target variables, unsupervised learning for discovering all the probabilistic relations that hold in your data, Clustering for new concept discovery
• Data importation Wizards allowing data preprocessing
• Missing values and hidden variables processing
• Evaluation of the models with confusion matrices, lift and ROC curves
• Analysis toolbox for easy models' understanding: strength and type of the probabilistic relations, sensitivity analysis, causal analysis, contradiction analysis, influence path analysis, HTML reports generation
• Powerful automatic layout algorithms for both tree networks and highly connected and complex graphs, Mutual Information Map
• High interoperability, with JDBC/ODBC connections to data bases, SQL interface, and exportation of Bayesian networks, tables, equations, graphs, matrices and reports just by copying and pasting.
• Dynamic Bayesian networks for taking into account the temporal dimension
• Decision Aiding tools with Decision and Utility nodes, direct representation of action policies, and reinforcement learning algorithms for automatic discovering of policies for static as well as dynamic Bayesian networks

... and plenty of new features that make Decision Aiding easier.

This new version can be downloaded  for a new evaluation period of 30 days. A dynamic presentation, a tutorial and some application examples are also available (Modeling and simulation of complex systems, Risk analysis, Mining customer data bases, Intrusion detection, Text Mining, MicroArrays analysis and Health Trajectory analysis).