First off, I have very limited knowledge statistics-wise and am more of a coder. I was thrown into a large scale project and could use some guidance.

I have a large multivariate time series dataset that I'm working with, where I'm trying to predict when an error will occur. There are several errors that could occur, however, I'm am trying to predict when only a certain error A will occur. The model I'm looking into is Recurrent Neural Network.

The data looks something like this:

time    param1    param2    param3 ... param30000    |    error
0s      30.202    9.2102    0          .201120       |    NULL
1s      30.202    11.177    1          .249165       |    NULL
2s      30.202    12.293    0          .171295       |    NULL
1930s   23.246    10.372    1          .302009       |    A

Obviously I plan to reduce dimensionality, but how could I tackle this problem to predict when A would happen again? How should I be structure the data? Time lagging? Any hints/guidance would be appreciated greatly.


Your Answer

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

Browse other questions tagged or ask your own question.