Time Series Forecasting Right Metric

Question

I'm doing a time series forecasting using Exponential Weighted Moving Average, as a baseline model. I'm wondering on what would be the best metric to use if I have a set of percentage values.

Is it okay to use RMSE to assess model's performance?

Here's the output,

Thanks.

Answer 1

Yes, RMSE is a very suitable metric for you.

Follow the blogs on machinelearningmastery.com This guy has written some very good blogs about time-series predictions and you will learn a lot from them.

Example blog for time series forecasting:

https://machinelearningmastery.com/time-series-prediction-lstm-recurrent-neural-networks-python-keras/

Example blog for loss function selection:

https://machinelearningmastery.com/how-to-choose-loss-functions-when-training-deep-learning-neural-networks/