This application allows the user to generate networks for crime linkage situations with two offenders according to the method introduced in [de Zoete et al., 2017] .

In the Set parameters tab, the user can specify

* The number of crimes.

* The number of different types of evidence.

* How often each type of evidence is present in the different cases.

These factors, together, determine the graphical structure of the Bayesian network. To fill the conditional probability tables, the user should specify whether Suspect X and Suspect Y match with the different evidence types. This can also be done in the Set parameters tab. Lastly, the user can set the labels used for the different types of evidence, i.e. instead of "evidence_1__1" , the nodes will have labels like "fingermark_1" . Note that the default settings match with the presented example ( Situation 5 ) from [de Zoete et al., 2017] .

In the Set probabilities and plot network tab, the user can specify the random match probabilities associated with the different types of evidence. Together with setting the number of people in the potential offender population this sets all the necessary (conditional) probabilities.

The application generates the network associated with these settings in real time. A graphical representation of this network is plotted. The network can be downloaded (using the ) download network text/button as .net file. Software packages like AgenaRisk, Genie and Hugin can be used to examine the network, add extra nodes and make computations.


de Zoete, J., Sjerps, M., Meester, R. (2017). Evaluating evidence in linked crimes with multiple offenders. Science & Justice,

Number of crimes and evidence types

Crime 1

Evidence types

Suspect X

Suspect Y

Generate BNs for two offender crime linkage scenarios

Download network