3
$\begingroup$

Let's say I'm doing some predictive analytics and am trying to predict US GDP per month using a two or three month lag. After every month, I generate new predictions and am able to compare my prediction to the actual gap. What would be the best way to represent this via plot. Line chart?

In all the research I've done, there doesn't seem to be a 'standard' way to visually represent time series predictions compared to the actual values over time.

$\endgroup$
2
$\begingroup$

Typically line graphs are used to represent trends. Alternatively, if your primary goal is to display the accuracy of your model, and if you have enough data, you could do a histogram of the differences between predicted and actual.

$\endgroup$
1
$\begingroup$

An addition to the histogram of the errors, is a plot of the errors over time. With the histogram you can't check the correlation of the error terms. In case the error terms are somehow correlated, you know your model can be improved.

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.