BIF_Generator

Code for Exporting a learned First-Order Bayesian Network to a Bayesian Network in BIF format.

Input:

• A relational database
• A learned Bayesian network (using FactorBase).

Output: An .xml file in BIF format.

How to Use

You have used BIF_Generator if you run the FactorBase on your relational database. The Bif_'YOURDATABASENAME'.xml is created in FactorBase/jar folder. If you wish to run the BIF_Generator separately, please follow these steps.

1. Import data into your database and run FactorBase.

2. Clone the BIF_Generator repository by

   git clone https://github.com/sfu-cl-lab/BIF_Generator.git

3. Navigate to BIF_Generator/cfg folder and modify the config file subsetctcomputation.cfg by copying the same config file content that you used to run FactorBase.

4. Navigate to BIF_Generator/jar where BIF_Generator.jar is located.

   Export the learned Bayesian Network (BN) by

   java -jar BIF_Generator.jar.

   OR

   Navigate to parent directory BIF_Generator and compile the .java source code by

   javac -cp ".:./lib/*" BIF_Generator.java
java -cp ".:./lib/*" BIF_Generator

5. The output file is created in the folder Bif_Generator named Bif_<YOURDATABASENAME>.xml

Visualize the Bayesian Network

We have tested the BIF_Generator with the AIspace tool from UBC.

1. Download and run the tool

   Download tool (.jar format) and
   Run by java -jar bayes.jar

   OR

   Download tool (.exe format) and
   Run the .exe file

2. Go to File > Load from File and select the Bif_<YOURDATABASENAME>.xml in the folder Bif_Generator.

You will obtain a result like that shown in the Image .