Preface
This document will introduce you
to the main functions of BayesiaLab by taking you through a very practical
guided visit. This is not BayesiaLab Help file (the help system is directly
available from the main menu).
Note: This tutorial is in no way
a course on Bayesian network theory; nevertheless, if this kind of training
is required, don't hesitate to contact us at this address: info@bayesia.com.
You will also find supplementary
explanations in the program's online help file and a general presentation
of this product can be found on our site: www.bayesia.com.
The tutorial is organized as follows:
Chapter
1: Creating a Bayesian network
This chapter will show you how
to graphically create a Bayesian network representing a particular expertise,
and from that to define the probabilities and to associate costs to information
acquisition.
Chapter
2: Using a Bayesian network
In this chapter, the Bayesian network
developed in Chapter 1 is taken further using Bayesian inference. The chapter
also shows us how to automatically generate an adaptive questionnaire based
on a target and how to evaluate the quality of the Bayesian network.
Chapter
3: Extracting a Bayesian network from a database
Data Mining methods can be used
to automatically generate a Bayesian network from the existing data. This
chapter will guide you in the use of the various automatic learning methods
found in BayesiaLab as well as for the use of tuning tools.
Chapter
4: Analyzing a Bayesian network
In this chapter, we will describe
the different tools proposed by BayesiaLab to greatly improve the readability
and the comprehensibility of the Bayesian networks: automatic positioning
of the nodes, sensibility analysis tools, and analysis report generation.
Chapter
5: Creating a dynamic Bayesian network
The Bayesian networks described
in the first few chapters are static Bayesian networks and they do not represent
a state of knowledge that is dependent on time. This last chapter illustrates
the modeling of a dynamic system, one that evolves over time. This modeling
uses a class of models called dynamic Bayesian networks.
Chapter
6: Learning policies
Bayesian networks are by themselves
decision aiding tools. However, this aid can be greatly improved. This chapter
will show you how using the Utility nodes for qualifying your states, how
defining Decision and lastly how learning automatically policies for your
static and dynamic Bayesian networks.
Chapter
7: Using Constraint nodes
This last chapter is dedicated
to the illustration of the use of the constraint nodes. These special nodes
allow specifying constraints that must hold between variables.
BayesiaLab
Tutorial