CODEWITHSUNDEEP
python-3.xforecastingevaluationarima
Logica
Logica

Reputation: 975

How to evaluate forecasted result by ARIMA model in Python

Click on the link to check the forecasted result of ARIMA I have dataset of daily basis having 1096 rows i have forecasted next 14 days using ARIMA Model. My question is that the forecasted result i got from ARIMA are actually unseen. Is there any evaluation metrics or any way to check whether the forecasted values that i got from ARIMA are right or wrong?

Is there any way to evaluate the unseen result?

Upvotes: 0

Views: 2589

Answers (1)

Steffen Jensen
Steffen Jensen

Reputation: 58

There are many metrics to evaluate forecasts, such as MAPE, MAE, ME, MPE, MASE, ...

For evaluating accuracy an absolute measure of the error such as MAPE or MAE can work well. Note, that all measures have each their purpose, as well as strenghts/weaknesses. Here two resources that can be used to get started, begin with the first and move to the second when you want the more detailed explanation:

Chapter 3.4 from Forecasting: Principles and Practice. Covers both fundamentals and implementation in R

"Another look at measures for forecast accuracy" by Hyndman and Koehler. Bit of discussion on the different measures, and suggestion of using MASE for evaluating against a naïve approach

Upvotes: 1

Related Questions