Child pages
  • WebSimulator Editor (6.0)

Contents

Context

Tools | WebSimulator Editor

The WebSimulator is a platform for publishing models via the web, which means that any Bayesian network model built with BayesiaLab can be shared privately with clients or publicly with a broader audience. The type of each variable and its associated graphical component is set in BayesiaLab by using the WebSimulator Editor. Once a model is published via the WebSimulator, end users can try out scenarios and examine the dynamics of that model. 

New Feature: Inputs | Mean Values

Prior to this new release, the mean values were only computed by using the MinXEnt. However, this type of evidence is used when you want to simulate the impact of a change in the population. 

When you are interested in using a continuous mean value to characterize one individual and not a population, the Binary Mean has to be used instead of the MinXEnt. This type of evidence is similar to what is done in fuzzy logic; the continuous value is created by interpolating between two adjacent states. 

 Monitor in BayesiaLabMonitor in WebSimulator

 

 

Unobserved

Soft Evidence

Mean with MinXEnt

Fixed Probabilities

 Monitor in BayesiaLabMonitor in WebSimulator

 

 

Unobserved

Soft Evidence

Mean with MinXEnt

Fixed Probabilities

 Monitor in BayesiaLabMonitor in WebSimulator

 

 

Unobserved

Soft Evidence

Mean with MinXEnt

Fixed Probabilities

 Monitor in BayesiaLabMonitor in WebSimulator

 

 

Unobserved

Soft Evidence

Mean with MinXEnt

Fixed Probabilities

New Feature: Inputs | Discrete States

Beside the Switches, BayesiaLab 6.0 comes now with a new component for setting evidence on discrete states: the Dropdown List. This is particularly useful when the variables have lots of states. 

 

New Feature: Outputs | Mean Gauges

Two gauges are now available for the outputs of mean values.

 

New Feature: Outputs | Function

We have added one component for returning the values of the Function nodes.

New Feature: Questionnaire Targets

In addition to the Simulators, the BayesiaLab WebSimulator offers now a new mode, the Adaptive Questionnaires

In an Adaptive Questionnaire, the variables/questions are dynamically (i.e. after each new piece of evidence) sorted according to their cost and the information they bring to the Target variables.

A new field is thus now available in the WebSimulator Editor to define the Target variables that will be used in the Adaptive Questionnaire.

Unlike the BayesiaLab's Adaptive Questionnaire, the WebSimulator one can handle more than one Target variable.