1
$\begingroup$

I am hoping that there is some existing software for what I want to accomplish, as I'm not a big fan of reinventing the wheel.

In general, I would like a software package that can serve as a workflow that integrates network analysis, predictive analytics, and performance evaluation.

More specifically, information about a criminal network along with other relevant data would be used to predict monetary flows across the network. Law enforcement case data combined with information about network monetary flows would direct resource use in disrupting flows across the network. These law enforcement performance metrics along with metrics related to criminal networks would be used for strategic decision making. It would also be ideal to have dashboards showing some of the metrics of interest.

I have identified some potential software candidates: Dataiku DSS, RapidMiner, KNIME, Orange Data Mining, Watson, and Statistica. I am not too familiar with them and I wonder if something else may immediately come to mind for those more familiar with these applications.

Any direction is greatly appreciated.

-Ted

$\endgroup$
0
$\begingroup$

In Orange, you can do something like this:

enter image description here

This takes the network, which is already containing class you'd like to predict, then training (or testing) the learner in Test & Score and evaluating it in Confusion Matrix. Then you can see misclassifications directly in the network graph.

There are a bunch of other learners and evaluation methods available. A big plus is also interactive data exploration (see how you can input wrongly classified data into Network Explorer?). However, there's no dashboard available yet. We make do with opening several windows side by side.

That's just my 2¢ on Orange. I suggest you to at least try all of them and see which one works best for you. :)

| improve this answer | |
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.