2
$\begingroup$

I am having difficulty structuring my data and finding a machine learning technique to predict my outcome.

My data: I have a number of users with observations of a number of factors each year, each user having a separate 'time-series', and I want to determine the outcome for the next year.

User | Year | Factor1 | Factor2 | FactorN | Outcome 1 | 2015 | 1 | 2 | N | 1 1 | 2016 | 2 | 2 | N | 1 1 | 2017 | 3 | 4 | Y | 1 2 | 2015 | 3 | 4 | N | 1 2 | 2016 | 2 | 3 | Y | 0 2 | 2017 | 1 | 4 | Y | 1 3 | 2015 | 3 | 4 | N | 1 3 | 2016 | 2 | 3 | N | 0 3 | 2017 | 1 | 4 | Y | 0

From my research I thought a recurrent neural network might be the right choice due to to its efficacy with sequences of data, but I'm unsure how to input the data I have into a RNN. I'm not even sure I'm tackling the problem correctly.

Is an RNN the right choice? Would it be more effective to 'flatten' the data into one row containing all years data or deltas between each year?

Thank you in advance for any advice.

$\endgroup$
  • $\begingroup$ You have longitudinal data with binary outcome so why don't you try some random-effects model with logistic regression? $\endgroup$ – user3119750 Oct 18 '17 at 18:42

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.