This forum is about

Next event prediction streaming learning

Posted by:
**
lv1984
**

Date: March 14, 2018 01:35AM

Hi,

I'm looking for a library/algorithm to predict the next event for a event sequence.

I saw that there are a lot of algorithm and I think that the library is suitable for me.

But I need something particular, I need to train the algorithm incrementally, when for example an end event arrives.

Example:

A->B->C->KO, send the sequence to the algorithm as training sequence

D->E->F->OK, send the sequence to the algorithm as training sequence

and

G->H, send the sequence to the algorithm to predict the next event.

Can I use the algorithm in spmf in this mode?

Specifically the sequence prediction algorithms?

Thank you

I'm looking for a library/algorithm to predict the next event for a event sequence.

I saw that there are a lot of algorithm and I think that the library is suitable for me.

But I need something particular, I need to train the algorithm incrementally, when for example an end event arrives.

Example:

A->B->C->KO, send the sequence to the algorithm as training sequence

D->E->F->OK, send the sequence to the algorithm as training sequence

and

G->H, send the sequence to the algorithm to predict the next event.

Can I use the algorithm in spmf in this mode?

Specifically the sequence prediction algorithms?

Thank you

Posted by:
**
webmasterphilfv
**

Date: March 14, 2018 07:40AM

Hi, yes, in theory, all the sequence prediction models provided in SPMF are updated incrementally by inserting one sequence by one sequence one after the other, and you can ask for a prediction for a new sequence at any time.

So I think that it meets your requirement.

Best regards,

So I think that it meets your requirement.

Best regards,