2
$\begingroup$

I need to create an app to predict events.

I'll explain my requirements: I have a set of intermediate events (eg.: 1, 2, 3...9) and a set of final events (eg.: A for OK and B for KO). These events bring a correlation id to relate each other.

What ML algorithm can I use to predict if a sequence of events could cause a future KO event (B in my example) or not?

I need an unsupervised algorithm.

Any idea about the algorithm?

$\endgroup$
0
$\begingroup$

Maybe sequential association rules could be applied here. Please check spmf java library, section 'Sequential Rule Mining'.

$\endgroup$
0
0
$\begingroup$

Start with a simple clustering algorithm such as k-Means?

Another idea is "Isolation Forest" on SK-Learn. Rather than clustering, you're looking for anomalies in your dataset and classify accordingly.

Good luck!

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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