BayesiaLab

An unequalled panel of functions

Network modelling: capitalize on your expertise and quantify your uncertainties

Thanks to its ergonomic interface, BayesiaLab enables you to easily formalize your knowledge in the form of bayesian networks.

• Highly intuitive graphic development of networks: single click creation of nodes and “draw arcs”, arc inversion functions, magnetic grid and alignment tools helping you to position nodes effectively
• Powerful equations editor for a quick description and concise relations (discrete and continuous probability distributions, arithmetic and trigonometric functions, ordinary operators)
• Full copy/paste functions (clipboard system interoperability) and an import function for the main formats on the market

Learning/data mining: discover the knowledge buried in your data

• Data import wizard allowing you to import from any source (text file, all the best known databases on the market)
• Ability to filter variables or unused values, to discretize continuous variables, to incorporate discrete modalities
• Very wide range of learning algorithms

Click here for more information on the data mining functions of BayesiaLab »

Analyse your models at a glance

BayesiaLab makes it easier to understand your models, thanks to its wide range of varied analysis tools and its automatic node positioning algorithms. This means that you can discover the knowledge buried in your data by visually analysing the Bayesian networks obtained and through using the toolbox:

• Arc analysis: width related to the strength of the probabilistic relation
• Analysis centred on a target value: characterization of the strength and the type of node relation in relation to the target modality
• Target node modality analysis allows you to visualize two types of information for each node linked to its probabilistic relation with the target: the type of influence a variable has on a particular modality of the target variable, the gain in information provided by the node on the knowledge of the target modality
• Causal analysis through the intermediary of important graphs
• Analysis of contradictions in the set of observations
• Generation of HTML reports on full target node analysis

Exploit the power of Bayesian networks

• Take advantage of the bayesian power of inference to simulate scenarios or take symptoms into consideration in order to diagnose your system
• Associate the acquisition costs of the variable values and BayesiaLab will dynamically generate adapted questionnaires suggesting the variables that need to be entered to obtain more information of the target variable at a lower cost

  

• To make the deployment of your bayesian scoring functions easier, export the Markov blanket of your target variable into an external programming language (SAS, PHP, etc.)
• Represent, assess and learn your policies thanks to:
  • 
  • Decision nodes to represent your Actions
  • Quality tables associated with actions in order to specify policies
  • Utility nodes for the qualification des reports
  • Learning by reinforcement algorithms for the automatic acquisition of policies allow the optimisation of the sum hoped for from the Utilities (finding optimal action policies)

Introduce the temporal dimension

Thanks to its temporal arcs, BayesiaLab allows you to integrate the notion of time into your networks and to model the dynamics of your systems.

• Follow the temporal change in the probability distributions of your variables, step by step, on your monitors.
• Start a simulation over a defined period of time, specifying the status of the variables to be followed temporally, on a graphic or in a file.
• Measure the impact of your preventive maintenance and risk control policies in terms of winnings/costs hoped for during the period of your choice.

See the general presentation of BayesiaLab animation (5 min.)  or visit our resources page to see examples of applications.

And to download a trial version of BayesiaLab, click here! »