I have a simple feed forward neural network regression model that I'm training on customer data to predict their usage amount. The MAPE is above 50%. The data is heavily skewed and when I log transform the predictor set as well as response variable the MAPE comes down to 12%, however reverse transforming the predicted value to original scale blows up the MAPE again.

The predictor set is close to 80 variables which has a good representation of customer behavior. However, Pareto NBD model predicts at 34% MAPE with just 4 variables. The MAPE computed for Pareto NBD is at Decile level. Does training Neural network at decile level makes sense?

Highly appreciate inputs.

  • $\begingroup$ It makes sense if your objective is to predict deciles. But, you're not asking this to get permission to do so, so what exactly is the broader scope of your question? It sounds like you're trying to boost your model's accuracy? $\endgroup$
    – Alex R.
    Mar 8, 2019 at 19:02
  • $\begingroup$ @AlexR. - Yes I'm looking to boost the model accuracy. I'm surprised how Pareto NBD is performing so well with just 4 variables where as neural network is not. Any thoughts? $\endgroup$
    – iprof0214
    Mar 12, 2019 at 21:11


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.