Brute Force recognition

Question

Hello machine learning community, 

We try to make a POC using machine learning in a small time frame in my company (kind of challenge game we have). 

During these three days, we would like to discover and experiment machine learning around a topic that is: "detect brute force attacks".

Our example is to imagine a system where you have to enter a digit code 6 to 9 digits to enter.

You can try three times and if all failed you have to reconnect.

We create multiple data set with 

trya, datetime, ishackingScenario, digitx, digity, digitz.

We create dataset with brute force hacking (+1 meaning digity=digitx+1 and digitz=digity+1). with +10 and other we imagined

We create some other dataset with valid connection or with normal errors that we can imagine.

So we are ready to train the system and we are able to tell when this is a problem and when this is not.

Then we would like to be able to send a csv with a mix of valid and invalid lines and we would like to see if the system is able to recognize a brute force pattern and alert us.

Any idea of the direction we should use to try to achieve this?

Any inspiring tutorial we should look at?

Any help is welcome.

Thank you all for reading till there.

Richard.

Answer 1

Hi Richard,

that seems funny. As far as I understand the observations in your dataset are labelled, that is, for training purposes you know in which cases the try was an attack or a valid connection. So the machine learning model should be a supervisioned one. Moreover, as the labels are two categories, you should apply a classification model. There are lot of two-class classification algorithms in AML Studio, you can train several of them and compare the scoring performance on your test dataset. I would try with a neural network applying hyperparameter tuning, given the pattern recognition task you need to conduct.

Best,

Ariel

Answer 2

Hello Ariel.

Thank you for your advice. We will look at these models.

I will give you some feedback on the final result of our game.

Thanks 

Richard.

Answer 3

Of course! I'm very curious about the results.